CHILDREN'S KNOWLEDGE AND MEMORY
CORRELATION BUT NOT CAUSALITY
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
lewis and charlie,
loves of my life
Many people helped to insure the successful completion
of this study. I thank Pat Miller for guiding me through the
intense graduate school learning process. Pat involved me in
her research early in my graduate school career, and her
enthusiasm for research was contagious! Pat always makes
time when her students need assistance and is a very
dedicated professional. Pat also showed all of us that a
woman can have a serious career and a family simultaneously.
She not only was my adviser but a terrific friend as well.
I would like to thank two members of my committee, Scott
Miller and James Algina, who taught me about research design
and statistics, both in and out of the classroom. Both were
excellent teachers and assisted me with numerous questions
throughout my graduate school career. I also thank the other
members of my committee, Ira Fischler, Rich Griggs, and Robin
West, for their support.
I am grateful to Peggy Blitch and the teachers and
students at Martha Manson Academy for giving me and my
assistants a warm welcome at the school. I also appreciate
the numerous occasions that Ruth Duncan and the teachers and
students at P.K. Yonge Laboratory School did the same.
Two fellow graduate students rescued me from doom, when
I thought making the videotapes would be impossible. Al
Avendano had the "know-how" to film and produce them. Pat
Aloise narrated them, because I lost my voice on the day the
camera was set to roll! The cooperation among graduate
students truly makes the program at the University of Florida
I was fortunate to have the assistance of several very
dedicated undergraduate assistants, who contributed many
hours of their time and talents to help make this study
possible. Carol Lewald and Rick Anderson helped to test
subjects for the study. Kathleen G. Cossey and Carol Lewald
coded and scored (and scored and coded!) the data. Others
who assisted with various aspects of the study were Randi
Clement, Troy Zarcone, Jackie Spiegel, Spencer White, Nikarre
Redcoff, and Coreen E. Prochnow at the University of Florida;
Lisa Mallet, Jean Muransky, Rose Kramer, Gary Kolvachik, Beth
Owen, and Teri Miller at Muskingum College; and Erica Miller,
a Sarah Lawrence student-to-be.
Cecil Chapman typed the labels for the birds on, what
was then, her brand-new laser printer. Cecil helps everyone,
and her cheerfulness makes the developmental/cognition area a
brighter place to be.
Bill Hardy was the much-depended-upon ornithologist who
helped with the original lists of birds, loaned me slides for
the videotapes, and answered, what seemed to be, my 10
million questions about birds. I knew nothing about birds
when I began this project!
I thank Larry Normansell, my colleague at Muskingum
College, for his assistance in producing a poster version of
this paper. The poster for the Conference on Human
Development in Charleston was aesthetically pleasing!
I thank David Bjorklund for keeping me informed of his
research and sending me a computer program for defining which
children were "strategic." I became excited about conducting
memory development research after attending a symposium at
the Conference on Human Development in Nashville (1985)
during which David, Peter Ornstein, and others debated the
nature of strategies.
I especially thank Lewis for enabling me to pursue a
career. In addition to his generous donation of time
assisting with computer questions and problems, Lewis is
always helping behind the scenes . I appreciate all that he
Thanks to my wonderful graduate school friends of the
past: Gillian, Marite, Malia, Vern, John, Terry, Nan, and
Carol for providing "kinship." The spirit and warmth of
Florida live on in all of us (though most of us have migrated
north) . Now as another clan of graduate students makes their
way up the graduate school rungs, may they keep the spirit of
Florida alive: Pat, Maggie, Janet, Nanci, Bob, Al, Wendy,
and others. Until we meet again, see you at SRCD!
TABLE OF CONTENTS
LIST OF TABLES viii
The Beginning of Research Related
to Memory 1
Review of the Literature Investigating
the Answer to "What is Memory the
Development of? 2
Strategy Use Develops 3
Knowledge Develops 8
Current Debates in the Field of Memory
The Definition and Measurement of
Purpose of this Study 19
Subj ects 25
Memory Test 3 2
Knowledge Tests 3 3
Coding and Scoring 36
Memory Test 3 6
Knowledge Tests 37
Overt Study Strategies 39
Did Knowledge Change? 42
Within-Subj ect Analyses 42
Between-Subject Analyses 49
Did Recall, Clustering, and Strategy Use
Within-Subj ect Analyses 52
Between-Subj ect Analyses 54
Individual-Subject Analysis of Changes in
Knowledge and Memory 56
Were There Significant Correlations Among
Knowledge, Clustering, Strategy Use,
and Memo r y ? 5 8
Summary of the Results 6 2
Why Correlation But Not Causality? 66
Other Issues and Future Research 74
A. LISTS OF BIRDS USED FOR THE MEMORY TEST 7 9
B. ORDER OF LESSONS PRESENTED ON THE VIDEOTAPES
ABOUT BIRDS 80
C. SAMPLE LESSONS FROM THE VIDEOTAPES ABOUT
D. LISTS OF BIRDS FOR THE FACTS TEST 8 3
E. SCORING FOR THE MATCHING TEST 84
F. SCORING FOR THE CATEGORY LISTING TEST 85
G. FORMULA FOR THE ADJUSTED RATIO OF CLUSTERING
(ARC) MEASURE 86
H. STUDY STRATEGY CATEGORIES AND THE BEHAVIORAL
REFERENCES 8 8
BIOGRAPHICAL SKETCH 94
LIST OF TABLES
2-1 Testing Sequence for Three Groups of Children
Tested for their Recall and/or Knowledge of
3-1 Pre-VIDEO and Post-VIDEO Means (and Standard
Deviations) for Each Knowledge Test 42
3-2 Means (and Standard Deviations) of Pretest
and Posttest Recall for Males and Females
in the Memory Test Twice Group 5 3
3-3 Means (and Standard Deviations) of Pre-VIDEO
and Post-VIDEO Clustering and Recall for
Younger and Older Children 55
3-4 Individual-Subject Analysis for Patterns of
Changes on the Knowledge and Memory Tests
by Age 57
3-5 Post-VIDEO Correlations Between Knowledge,
Clustering, or Strategy Use and Recall for
Groups of Children 59
3-6 Correlations Among the Knowledge Tests
by Age 60
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
CHILDREN'S KNOWLEDGE AND MEMORY:
CORRELATION BUT NOT CAUSALITY
December 19 88
Chairperson: Patricia H. Miller
Major Department: Psychology
Previous research on memory development suggests that
increases in knowledge during development may encourage
strategy use and improve recall. Unlike these correlational
studies, the present study tested this hypothesis by
experimentally manipulating knowledge.
Knowledge about birds among 44 younger (mean age: 8-7)
and 53 older (mean age: 10-5) children was assessed via four
tests. 1) GROUPING: Children placed 45 picture/name cards of
birds into groups that they thought belonged together, and
titled each group. The number of birds placed into meaning-
based groups (e.g., pet birds) was tallied. 2) LISTING:
Children said all the members they knew for six experimenter-
designated categories (e.g., birds of prey, songbirds). 3)
FACTS: Children stated all the facts they knew about each of
the birds. 4) MATCHING: Children placed the birds into the
six possible categories used for LISTING.
Children's strategies while studying 20 picture/name
cards of birds were videotaped, coded, and scored. Depending
upon their assigned groups, children took either the four
knowledge tests or a memory test both before and after, or
they took all tests only after viewing five videotapes about
birds. The videotapes, shown in the classroom by the regular
classroom teacher, taught information about the birds and
emphasized intra-group associations.
Between-subject analyses showed significantly higher
knowledge scores after the videotapes than before for both
sexes on MATCHING, and for boys only on GROUPING. Within-
subject analyses revealed significant improvements on
MATCHING, LISTING, and GROUPING. All analyses showed
substantial increases in the percentage of facts referring to
general, categorical information.
Despite significant increases in knowledge and
significant correlations between knowledge and memory, there
were not significant improvements in clustering and overt
study strategies (e.g., rehearsal). Although knowledge
changes were greater overall for males, only the females in
the within-subject analysis showed significant improvement in
recall. Males' recall decreased significantly. The between-
subject analysis showed no improvement in recall for any
Either knowledge is not the cause of improvements in
memory with age, or knowledge is necessary as a first step,
but not sufficient. Possible reasons for correlation but not
causality are discussed.
The Beginning of Research Related to Memory
The study of memory is a topic that has intrigued
psychologists for more than 75 years. One of the earliest,
and most famous, studies of memory was conducted by Hermann
Ebbinghaus, a German psychologist. Ebbinghaus invented the
nonsense syllable and used himself as a subject. His measure
of memory was that of relearning, the time it took to relearn
a list of nonsense syllables to perfection, after various
intervals of time (Ebbinghaus, 1913).
Although other studies appeared during the first 60
years of the 20th century, that period has been described as
a "long period of misdirection, confusion, and sterility"
(Olson, 1983). However, the study of memory blossomed in the
late 60s, and new methodology facilitated its study.
Introspection, which was dominant in this country during
Ebbinghaus' time, was no longer considered a scientific
method. The information processing approach has prevailed
since the 1960s.
Research related to the development of memory
flourished following a landmark symposium, organized by John
Flavell, which took place at the biennial meeting of the
Society for Research in Child Development (SRCD). The title
of this 1971 symposium was "What is memory the development
of?" At the 1981 SRCD symposium which marked the ten-year
anniversary of Flavell ' s symposium, Trabasso (1983) remarked
that the answer to the general question "What is memory the
development of?" depends on one's answer to "What is memory?"
Our conceptions of what memory is have changed since 1971,
and so have our conceptions of what is developing, but no one
to date has been able to answer Flavell' s challenging
question with certainty.
Review of the Literature Investigating the Answer to:
"What Is Memory the Development of?"
Over the past 20 years, several answers to the question
"What is memory the development of?" have appeared in the
literature. One answer was that capacity developed (Pascual-
Leone, 1970). However, that answer has been challenged in
more recent reviews of the literature (Dempster, 1981, 1985)
and will not be discussed in this paper. Another answer
offered is that metamemory (i.e., children's understanding of
the memory process) develops (Pressley, Borkowski, &
O'Sullivan, 1985; Rao & Moely, 1987). That line of research
also will not be examined in this paper, because it is not
directly related to the present study. Two other answers to
the question have received the most attention. They are that
strategy use develops and that knowledge develops. Each of
these will be discussed.
Strategy Use Develops
One of the answers to the question "What is memory the
development of?" was emphasized by several presenters at the
1971 symposium and still is being researched today. That
answer is that strategy use develops. Theoretical models
from the 1970s and up through the early 1980s (if not
longer) also emphasized strategies (Ackerman, 1987).
In 1970 an influential article by Flavell proposed that
younger children have a production deficiency. That is,
younger children typically do not employ strategies for
remembering (e.g., rehearsal) spontaneously. However, if
younger children are instructed to use the strategy, their
recall rises to the same level as older children, who
spontaneously produce the strategy.
A surge of research investigated that hypothesis.
Improvements with age in strategies such as rehearsal (e.g.,
Flavell, Beach, & Chinsky, 1966; Ornstein, Naus , & Liberty,
1975); clustering (e.g., Moely, Olson, Halwes, & Flavell,
1969; Moely & Shapiro, 1971); organization (e.g., Corsale &
Ornstein, 1980; Kee & Bell, 1981); and elaboration (e.g.,
Pressley, 1982; Pressley & Ross, 1984) are well documented.
Training studies also reliably show that younger children
benefit by brief instruction in rehearsal (e.g., Naus,
Ornstein, & Aivano, 1977; Ornstein, Naus, & Stone, 1977);
clustering (e.g., Moely et al . , 1969); organization (e.g.,
Bjorklund, Ornstein, & Haig, 1977; Moely & Jeffrey, 1974);
and elaboration (e.g., Pressley et al . , 1985).
An example of a study from this period follows. This
study by Moely et al. (1969) was interpreted as confirming
that younger children have a production deficiency. Children
in grades 1, 3, and 5 were presented with 16, 20, or 24
pictures of categorically related items (i.e., 4, 5, or 6
items from the categories of animals, furniture, vehicles, or
clothing) to study for two minutes. There was greater
categorical organization with increasing age. Although there
was little change between grades 1 and 3 in the children's
study and recall organization, there was a sharp increase in
both measures between grades 3 and 5 . An instructions
condition, during which children were asked to sort the items
"that go together or are alike" (p. 28), facilitated the
performance of all children. Furthermore, after recall, when
nonsorters (children who had not been instructed to sort)
were asked to sort the pictures, 58%, 74%, 88%, and 95% of
the kindergarten, first, third, and fifth graders,
respectively, were able to sort the cards according to their
category membership. Thus, nonsorters were able to sort.
The fact that these young children were able to sort cards
according to category membership was interpreted to mean that
they possessed the necessary knowledge about categories, but
simply did not use it.
Much like the Moely et al. study, the other studies of
this period seemed to confirm that although younger children
(i.e., children in grades 1 to 3 ) often did not utilize the
appropriate strategies spontaneously, they easily could be
taught to use them, and consequently, benefited from them.
By grades 5 to 7 , children employed strategies spontaneously
More recently, several aspects of the production
deficiency hypothesis have been questioned. Chi and Ceci
(1987) challenged the assumption that young children have the
relevant knowledge about categories but simply do not use it.
Although the young children in the Moely et al. study (as
well as other studies of that period) were able to
successfully complete sorting tasks, and the researchers
claimed this was evidence that the children had the necessary
knowledge, Chi and Ceci suggested that children's knowledge
simply may not be in a readily usable form. They gave the
example that picking pictures that belong to the category
"animals" only requires a child to know whether the
attribute of "animalness" goes with each item. They believe
that this type of knowledge is different from the knowledge
needed to link elements to a unified superordinate structure.
That is, a child who can sort or match may have a
representation that is not a categorical structure.
Therefore, the child's and adult's representations of
knowledge still might differ dramatically with respect to
organization, and these differences might make the knowledge
less available to children.
Rabinowitz (1984) has argued for the importance of
differentiating between whether young children do not use
available knowledge because they prefer not to use it or
because it is less accessible to them. Although the
traditional notion of production deficiency claimed that
young children were predisposed to use strategies (such as
repeating a word over and over) that were less effective in
helping them recall, Rabinowitz suggested that those
strategies might be more easily accessed than one of
organizing words by semantic meaning. He demonstrated that
children use categorical processing more effectively when
the list contains highly representative items from the
A second question concerns the way in which strategies
develop. Several studies stimulated by the production
deficiency hypothesis have demonstrated what appear to be
strategic behaviors by preschoolers (e.g., Baker-Ward,
Ornstein, & Holden, 1984; DeLoache, Cassidy, & Brown, 1985;
Wellman, in press). This might seem to challenge the claim
that young children do not spontaneously produce strategies.
However, these occur only in specific, supportive contexts.
It is not clear how and why early strategic behavior emerges
(see also a review by Daehler & Greco, 1985), and how it
relates to mature strategy use, which appears to be more
general. Researchers currently are debating whether
earlier strategies are deliberate and intentional or
relatively unconscious and automatic. This and other debates
in the field will be addressed in a later section.
Even if younger children produce the same strategy as
older children, it may be, as suggested in a study by
Guttentag (1984), that strategy use is more effortful for
younger children. This may account for the observed lags in
the facilitative effects of early strategy use for younger
children (e.g., DeMarie-Dreblow & Miller, 1988; Miller,
Haynes, DeMarie-Dreblow, & Woody-Ramsey, 1986). These
children spontaneously produce an appropriate strategy, but
it does not help them, a phenomenon that Patricia Miller has
labeled a "utilization deficiency" (Miller & Harris, in
press). If producing a strategy is quite effortful for young
children, there may be little capacity remaining for
subsequent processing of the material to be remembered.
Thus, as these studies illustrate, clarifications of the
mechanisms by which strategies develop are necessary.
Although debates exist as to the development and nature
of strategies, probably no one in the late 1980s still
believes that strategy use alone can explain what develops
when memory develops. Another view of memory development,
which is receiving greater and greater attention from
researchers, suggests that knowledge plays an important role
in memory development.
A study by Chi in 1978 posed a major challenge to the
field of memory development research. She showed that when
younger children had more knowledge of a domain (e.g.,
knowledge of chess) than adults, the superiority of the older
subjects, typically found in other studies, was absent. In
fact, the 10-year-old children in Chi ■ s study (who were chess
experts) recalled more chess piece positions from immediate
memory and took fewer trials to recall all chess piece
positions, than the adults (who were chess novices). The
children also retrieved a greater number of chunks than the
adults for the first trial. On the other hand, Chi found the
more typical adult superiority for a digit span task. She
proposed that knowledge of the specific content area was more
important than general strategies in determining memory
performance. Thus, Chi ' s answer to the question "What is
memory the development of?" was that knowledge develops.
A study by Lindberg (1980) found similar results. Third
graders and college students were tested for recall of two
different lists, each containing 30 words. One list was a
typical experimental list that contained words from Battig
and Montague category norms. The other list contained words
generated by three third graders. These words were assumed to
be highly familiar to third graders. An example was the
members of the television show "Charlie's Angels," something
of importance to most third graders in 1980! Although the
college students clustered and recalled more of the words
typically used in experiments, the third graders clustered
and recalled more of the words that were more salient to
them. Again, a reversal of the traditional superiority of
adults seemed to show that knowledge of the stimuli was an
important aspect of memory development.
This wave of research in the late 70s and thus far in
the 80s has attempted to define the role of knowledge in
memory performance. Differences in knowledge were said to be
responsible for both differences in the production of a
strategy (e.g., rehearsal) and memory per se in a study by
Naus and Ornstein (cited by Zembar & Naus, 1985a). Naus and
Ornstein tested two groups of adults (soccer experts and
soccer novices) for recall of two different lists of words (a
typical experimental list and a list containing soccer-
related items). Although there were no significant
differences between the two groups for recall and size of the
rehearsal set (i.e., the average number of different words
rehearsed aloud per inter-item interval) for the typical
experimental list, the soccer experts were superior on both
measures for the soccer-related item list. Because adults
presumably already have sophisticated memory strategies
accessible to them, the results were attributed to the
differences in knowledge of the materials for recall.
Likewise, a study by Zembar and Naus (1985b) found
differences in the level of strategy use and recall,
depending upon the materials. They gave third graders, sixth
graders, and adults lists that varied in familiarity to them
and found corresponding differences in rehearsal strategies
and recall. When third and sixth graders were studying a
list that was very familiar to them, they used a more
sophisticated rehearsal strategy (i.e., rehearsing a greater
number of different items together) than when they were
studying a list that was less familiar to them. An
interesting result in this study was that the third graders'
rehearsal and recall of the easiest list was almost identical
to the sixth graders' rehearsal and recall of the typical
experimental list. On the other hand, the sixth graders'
rehearsal and recall of an advanced adult list was as poor as
the third graders' rehearsal and recall of the typical
experimental list. Thus, these studies suggested that
knowledge of the materials to be learned could influence the
use of strategies, which then could influence memory
In the words of Zembar and Naus (1985a):
In a "short-term" fashion, the available empirical
evidence suggests that age-related changes in knowledge
can affect the ease with which strategies are employed
by modifying the actual content in knowledge as well as
increasing the ease by which information can be
accessed, (p. 7)
It has become obvious to researchers investigating
memory development in the 19 80s that no single, simple
answer to the question "What is memory the development of?"
is likely to be found. It appears to be more and more likely
that knowledge and strategies interact. Recently,
researchers have begun to debate several aspects of strategy
usage and knowledge. Because these provide a context for the
current study, they are explored in the next section.
Current Debates in the Field of Memory Development
Accompanying the important question of what is
developing are recent debates about various issues
concerning the nature of strategies and knowledge. One
debate concerns whether strategy use during the grade school
years is deliberate and intentional or automatic and
unconscious. Another debate concerns the nature of knowledge
and why greater knowledge may facilitate the use of
strategies. Discussions of both follow.
Currently, the main debate concerning the nature of
strategies and their use is between Peter Ornstein and David
Bjorklund. Ornstein, Naus, and others take the position that
strategy use develops much like skills develop (see Ornstein,
Baker-Ward, & Naus, in press; Zembar & Naus, 1985a). They
believe that the cognitive operations required for the
successful deployment of strategies gradually become more
routinized. As this happens, strategies require less and
less capacity with age, and, consequently, during development
they facilitate children's performance to a greater extent
when they are employed. Ornstein, Naus , and others also
distinguish between strategies at input (e.g., rehearsal and
organization) and strategies at output (e.g., clustering).
Finally, they claim that even preschoolers show strategic
behaviors when learning very familiar material, and that even
these behaviors may be intentional and deliberate.
Bjorklund, on the other hand, disagrees with the
position of Ornstein, Naus, and others. He believes that
strategy use, and particularly the clustering that typically
is observed in children's recall, is not deliberate until
early adolescence (see Bjorklund, 1985). He claims that the
clustering typically observed in younger children's recall is
due only to the automatic activation of associated words in
memory. This position also was taken by Lange (1973,1978),
who included lists that minimized the associativity of words
and found that younger children's clustering was no longer at
Other evidence for that position came from research by
Bjorklund and Zeman (1982,1983). When children in first,
third, and fifth grades were asked to recall their
classmates' names (information that is quite familiar to
them), memory performance was high for children of all ages.
Although clustering (i.e., organizing recall by seating in
the classroom, reading groups, race, or sex) also was a I a
high level, the majority of children were not able to state
what strategies they had used to remember the information.
This was taken as evidence that early clustering is based
upon associative relations, and that strategies later are
discovered when subjects notice patterns of their own
Consistent with Bjorklund's position is the finding that
grade school children, who appear to be clustering at above-
chance levels, in fact do so minimally. Frankel and Rollins
(1982) attempted to explain why children who are under eight
years of age who had above-chance clustering did not
necessarily have enhanced levels of recall. Whereas the
younger children's recall typically consisted of isolates or
of only two intracategory items, adults' recall consisted of
category strings, with as many as five or six items per
string. The results were interpreted as evidence that the
organization of children's and adults' recall differs
dramatically, and that children's clustering in recall may be
associative rather than categorical.
Bjorklund (1987) proposes that a major developmental
change is the increased frequency of occurrence of particular
items. He argues that more frequent exposure to certain
items probably causes an increase in the number and type of
features and creates more links with other items in semantic
memory. This, he believes, is largely responsible for the
changes with age in the ease with which semantic memory items
can be activated. In a more recent review of the literature,
Bjorklund and Muir (in press) further suggest that when items
in semantic memory can be activated more easily, children
have more mental capacity available for other cognitive
operations (e.g., memory strategies). Bjorklund (1987) also
cautions that one must be careful not to define the term
"knowledge" too broadly when referring to differences in
knowledge base across ages.
The Definition and Measurement of Knowledge
Although the focus of many current studies of memory
development is on differences in knowledge, very few have
given much attention to a) quantifying the knowledge or b)
comparing the recall of individuals with moderate differences
in knowledge. Many studies simply classify individuals as
experts or novices (Chi, 1978; Gobbo & Chi, 1986; Naus &
Ornstein, cited by Zembar & Naus, 1985a), neglecting the
relative degree of knowledge present. Although the knowledge
differences in which developmentalists are interested
primarily are those involving a moderate variation in type or
degree, most studies investigating knowledge effects define
knowledge in terms of only two points on the continuum,
typically very advanced or very elementary knowledge.
Saarnio (1987) claims that studies of this nature inflate
the role of knowledge. During the tenth anniversary
symposium of Flavell's 1971 symposium, Olson (198 3) noted
that there still were several important clarifications to be
made in the field of memory development, one of which was to
determine more subtle differences between experts and
novices, beyond the obvious one that the expert knows more
about the topic than the novice. There may be qualitative
changes (i.e., changes in organization) occurring that go
beyond the simple differences in amount of knowledge when
expertise reaches the highest level, but this cannot be
determined unless attempts first are made to quantify
Other studies compare words that are presumed to be more
familiar to one or more groups than others (Lindberg, 1980;
Zembar & Naus, 1985b) or compare recall for more typical
versus atypical members of categories (Bjorklund, 1988;
Bjorklund & Thompson, 1983; Buchanan & Bjorklund, 1988). All
of these studies measure knowledge at two points at the upper
end of the continuum, that is, knowledge that is extremely or
moderately well-learned. Again, knowledge has not been
measured along a continuum. Therefore, the challenge for
memory development researchers in the 1990s will be to
specify exactly what is meant by knowledge and exactly what
are the actual mechanisms by which memory improves . Very few
studies have attempted to examine the contents of knowledge
One study quantified the meaningfulness of three letter
nonsense syllables (trigrams) by having subjects generate as
many associations for each as possible. Differences in
recall also were measured. It was found that differences in
recall could be explained by differences in meaningfulness
(Richman, Nida, and Pittman, 1976). However, because the
recall test consisted of non-words, it is difficult to
generalize the results with certainty. Using a similar
measure but with real words (i.e., requiring adults to
generate as many facts as they knew for each of 45 birds),
DeMarie-Dreblow (1988) found a different outcome. Degree of
meaningfulness did not correlate significantly with adults'
recall for birds.
Chi and others conducted research specifically designed
to describe the structure of knowledge in memory. For
example, Chi and Koeske (1983) compared the knowledge
structures of two subsets of a knowledge domain that one
child possessed. They utilized the model of memory proposed
by Collins and Quillian (1969), which assumes that concepts
are stored in memory as nodes on an associative network, and
relations between concepts serve as associative links. Their
subject, a 4 1/2 year old boy, was an expert on the subject
of dinosaurs. They compared the way that familiar dinosaurs
were structured in his memory to the way that less familiar
dinosaurs were structured.
The first task entailed having the child generate all
the names of dinosaurs he could. He produced 4.6 distinct
names over a two week period. Next, the experimenter and
child switched roles in a clue game, during which one person
told a list of properties and the other had to identify the
dinosaur which possessed those properties. Then, two lists
of 20 dinosaurs were presented at the rate of one word every
three seconds for the recall task. Finally, one year later,
a recall naming task was presented to test his memory. The
child was shown pictures of dinosaurs and asked to name them.
Chi and Koeske's results showed that the dinosaurs that
were more familiar to the boy had a greater average number of
links associated with each one, and that these dinosaurs had
a greater number of intra-group links with very few, if any,
inter-group links. Whereas the boy recalled 10, 8, and 9
dinosaurs across trials from the more familiar list, he
recalled only 6, 4, and 3 dinosaurs from the less familiar
set. One year later, the boy named 11 of the 20 familiar
dinosaurs when shown picture cards, but named only 2 of the
20 less familiar dinosaurs correctly. Chi and Koeske
concluded that the observed differences were due to the
difference in the structure of knowledge in semantic memory.
They suggested that the knowledge of the more familiar set of
dinosaurs was more cohesive.
Although this study measured both knowledge and memory,
there was only one subject. Therefore, it is difficult to
generalize these results with any confidence to all children,
or to be certain that they explain memory development in
In another study, Gobbo and Chi (1986) utilized
production and sorting tasks to compare the underlying
knowledge structures of five experts and five novices in the
domain of dinosaurs. Expertise was determined via a pretest,
and all children were seven years old. For the production
task, children said the name of each pictured dinosaur and
told everything they knew about it. The number of explicit
propositions (i.e., propositions based solely on information
that could be gained from the picture) versus implicit
propositions (i.e., propositions that gave interpretations or
made reference to information that could not be directly
observed in the picture) was determined from the children's
protocols. Whereas the two groups did not differ in their
use of explicit propositions, the experts used a
significantly greater number of implicit propositions than
the novices. Furthermore, on the sorting task, experts were
less likely than novices to group the dinosaurs on the basis
of explicit physical features. Instead, experts tended to
group the dinosaurs on the basis of combinations of
attributes, resulting in family distinctions in many cases.
Gobbo and Chi emphasized that experts and novices differ
in their representations, particularly in that experts tend
to focus on "deep-level," rather than "surface-type,"
features. Experts' knowledge also tends to be more
structured and cohesive. Chi and Koeske suggested that these
differences in knowledge are responsible for the differences
that have been observed in the way experts and novices use
their knowledge. However, neither a memory test nor any
measurement of strategies was included in this study, so it
is difficult to relate these results to issues of memory
Although Chi (1985) interpreted the recall differences
in the Chi and Koeske study as resulting from domain-specific
strategies, Ornstein and Naus (1985) cautioned that the study
neither defined nor measured particular strategies.
Furthermore, they claimed that it is important to
differentiate the contents of knowledge from the ease with
which that knowledge can be retrieved from memory. Current
research does not differentiate between these two aspects of
a knowledge base. In fact, there have been no systematic
studies of relations among knowledge, strategy use, and
Purpose of this Study
Many aspects of this study differed from previous
studies of memory development. These differences were as
1. Previous research was correlational (i.e., examined
relations between memory and pre-existing differences in
knowledge). In contrast, this study examined the claim that
knowledge causes improvements in strategy use and memory by
experimentally manipulating knowledge (i.e., constructing new
knowledge). To the writer's knowledge, this is the first
experimental test of that hypothesis. Because correlation
need not imply causality, such a test is essential. After
pretests for knowledge of, or memory for, birds, knowledge
was manipulated by showing children videotapes about birds.
If the claims of previous research are correct, then
increases in knowledge should be accompanied by increases in
both study strategies and recall.
2. Previous research did not measure knowledge along a
continuum, but instead tended to select subjects who
represented the endpoints (i.e., experts and novices) of
knowledge. In contrast, this study measured moderate levels
3. Previous studies generally included only a single
measure of knowledge. However, as is clear from the debate
on the nature of knowledge described earlier, there may be
different aspects of knowledge, which may differ in their
effects on memory. Thus, this study included four different
knowledge tests, each of which quantified knowledge.
The first test was similar to the sorting task utilized
by Chi and Koeske (1983). Children were asked to group birds
that were alike or went together in some way. Based on Chi
and Koeske 's differentiation between explicit and implicit
propositions, this study included a count of the number of
birds that were placed into meaning-based groups (i.e.,
groups formed on the basis of information that could not be
obtained by information that appeared in the pictures of
birds ) .
The second knowledge test was similar to the measure of
production in Chi and Koeske's (198 3) study. Children were
asked to generate as many birds as possible that belonged to
each of six different experimenter-designated categories.
The total number of birds correctly generated by a child for
these categories was tallied.
The third knowledge test was similar to Richman, Nida,
and Pittman's (1976) measure of meaningfulness . Children
were asked to generate as many facts as they knew for each
bird. The total number of correct facts generated by a child
and the total number of birds for which a child knew at least
one fact were calculated. Then, the nature of the generated
facts was examined to determine whether a larger percentage
of the facts referred to more general, categorical
information after children watched the videotapes.
The fourth, and final, knowledge test was a matching
test. Children decided to which of six categories each bird
belonged. This test only required knowledge of which
category "attribute" best described each bird, so it was
similar to the measures of knowledge reported in early
research on children's use of organizational strategies
(e.g., Moely et al., 1969).
4. Although researchers claim that the structure of
knowledge may influence the way that knowledge is used,
again, no one has tested that hypothesis experimentally. For
example, Gobbo and Chi (1986) found that experts had more
intra-group and fewer inter-group connections in memory than
novices. They hypothesized that the fact that experts'
knowledge was more tightly structured might make that
knowledge more accessible to them.
The domain of birds was chosen because it seemed to be
an area about which everyone would know something. Yet,
knowledge of birds was expected to be highly variable, with
some individuals knowing quite a bit and others knowing very
little. Therefore, it appeared to be an area in which
knowledge would fall along a continuum. Also, birds can be
classified on a simple physical (e.g., size of beak) or more
conceptual basis, much like the dinosaurs were in the work of
Chi and others.
In this study, videotapes about birds were shown to
children to attempt to create knowledge that was more tightly
structured. Lessons about birds taught intra-group
associations (i.e., associations were made between that bird
and two other members of the bird's group) in an effort to
see if creating the stucture of that type of knowledge would
facilitate strategy usage.
Much as in Lange's studies (197 3, 1978), the
associativity of words on the recall lists was minimized by
not placing words commonly clustered by children and adults
in pilot research on the same list. Therefore, this would
reduce the words clustered due to their associativity and
enhance the possibility of detecting clustering due to
category relatedness. in turn, any changes in clustering
would more likely be due to the teaching about birds on the
videotapes. To insure that any improvements would not be due
only to experience with particular items, children received a
different list to recall on the pretest and posttest.
5. Most of the previous research focused upon knowledge,
or a single study strategy, and recall. To date, no
systematic study of all three (knowledge, strategy use, and
recall) has been conducted with a large number of subjects.
In this study, children in grades 2 through 5
participated. Previous research suggests that children begin
to use more active study strategies sometime during those
years. Children were videotaped during the study time for
the memory test, and several overt study strategies (i.e.,
rehearsal, self -testing, and involvement with the cards such
as touching or moving them) were identified. In addition, a
measure of clustering during recall was assessed. Like other
studies (e.g., Moely et al., 1969), the memory test was a
recall test. After children studied 20 picture cards of
birds for two minutes, they were asked to recall them in any
order. Thus, this study was comprehensive and included
measures of knowledge, memory, clustering, and study
There were three major hypotheses in this study. The
first hypothesis concerned the measurement of knowledge. It
was hypothesized that knowledge, as measured by each of the
four knowledge tests, would improve significantly after the
children watched the videotapes about birds.
Second, as mentioned previously, current theories
suggest that changes in knowledge may be responsible for
improvements in both strategy use and recall with age. If
this were the case, then improvements in knowledge should be
accompanied by more active use of study strategies, and
improvements in clustering and recall. If changes in
strategy use and recall were not found, that would question
whether there was a causal relation between knowledge and
Whether or not strategy use and recall improved
significantly, the third hypothesis was that there would be
positive correlations among knowledge, strategy use,
clustering, and recall. This would support previous
research. However, whether the relation among them would
change with age was not known.
Parent permission letters were sent home with 124
children attending a private school in Gainesville, Florida.
As suggested by the fact that the tuition for one year was in
excess of two-thousand dollars, the children generally were
from middle-or upper-middle class families. Nearly all were
Parents of 97 children consented to participation in the
study. There were 9 males and 12 females from second grade
(mean age: 8-1), 14 males and 9 females from third grade
(mean age: 9-2), 14 males and 18 females from fourth grade
(mean age: 10-1), and 11 males and 10 females from fifth
grade (mean age: 11-1).
Each classroom teacher rated her students' ability to
learn in school by assigning a descriptive number from one
(low) to three (high). Children within each grade, sex, and
achievement rating were assigned randomly to groups with the
constraint that each group contain children with
approximately equal mean ratings.
Picture cards were utilized for the memory and grouping
tests. Each of the 45 three by five inch cards contained a
black line-drawing of a bird in the center and a label with a
single word naming the bird at the bottom (e.g., "owl" for
great horned owl). Because all birds' pictures were black
and white and were drawn to be approximately the same size,
there were no size or color cues provided by the cards.
The picture cards for the memory test had a pale blue
background, and the naming-label appeared on a small, white
strip. The picture cards for the grouping test were xerox
copies of the others. Thus, these were only black and white.
Pilot testing of children in grades 3 to 6 at a
different school was conducted in order to develop two
different lists of 20 birds (see Appendix A) to be used for
the memory test. Based on the pilot data, these lists had
approximately the same probability of recall. In addition,
birds that tended to be clustered together by college
students (determined by other pilot testing) were placed on
different lists. With lists that contained few commonly
clustered words, the effect of instruction (giving additional
knowledge of birds) upon clustering could be examined more
clearly. Each list contained the same number of members
(i.e., three or four) that could be classified in each of the
six categories. The birds belonging to each of the six
categories were determined by an ornithologist at the Florida
Museum on the campus of the University of Florida. Thus,
children who took the memory test twice received different
(but approximately equal) lists of words, and any improvement
in recall would not simply be due to recalling the same list
Children's study strategies during the memory test were
videotaped with a Sony Camcorder attached to a tripod in a
corner of the room, approximately 7 feet from the table at
which children sat. There was no monitor. A stopwatch timed
both the study period and the time for computing math
problems. Children's responses during the memory test and
the facts test were tape recorded.
A yellow tagboard strip was utilized for the matching
test. This strip contained one representative picture for
each of the six categories and keywords for each category
label (e.g., a picture of waves and "live by the water" or a
picture of a place setting at a dinner table and "people
Five 10 to 15 minute videotapes (VIDEO) taught facts
about 34 of the 45 birds (see Appendix B). Only a subset of
the tested birds were taught, so that the other birds could
serve as a control set. Analysis of the control set would
help to reveal the source of changes in children's
performance on the tests. If no improvement occurred in the
control set, then improvements in the set of birds that were
taught probably were due only to the videotapes. However, if
improvements were noted on the control set, that might
suggest that taking the tests alone improved children's
A graduate student filmed pictures of slides with the
Sony Camcorder, while a female graduate student narrated the
lessons. Thus, these videotapes were of the "home video"
A brief, general introduction to birds was provided at
the beginning of the first videotape. A brief introduction
to each of the six categories was given before the first
member of that category was introduced. Thereafter, no two
birds from the same category were taught consecutively, and
there was no explicit teaching of all the birds belonging to
any specific category. Because this was the first
experimental investigation of the hypothesis that increases
in knowledge (via additional intra-group associations and a
greater number of connections per bird in semantic memory)
cause improvements in memory, it was important not to
explicitly teach the categorical structure. If children had
not been required to do any construction of knowledge
themselves, any improvements in memory would appear trivial.
The format for teaching about each bird was similar.
Each segment began by telling how that bird was similar to
the one in its category that most closely preceded it. Many
facts about the bird's habitat, nesting, migration, food,
mating, song, or peculiar habits were told during the lesson.
Similarities and differences between that bird and its two
category members that preceded it most closely were given
during the lesson, and the lesson ended by explaining
something unusual about that bird, giving a reminder about
the bird, or telling a funny story about that bird (see
Appendix C for sample lessons). The sources consulted for
information about birds were books (Bull & Farrand, 1977;
Peterson, 1947; Robbins , Bruun, & Zim, 1966; Steward, 1977;
Zim & Gabrielson, 1956) and an ornithologist.
Other cues were provided during the teaching of some of
the categories. Songbirds' songs were audible during entire
lessons on songbirds. Pictures of water birds always showed
the bird on, near, or over the water. Nightbirds' sounds
were played briefly sometime during the lesson, and either
their pictures had a dark sky, or the bird was shown
camouflaged among its natural surroundings.
Two groups took four knowledge tests or a memory test
both before and after viewing the videotapes about birds (see
Table 2-1). A third group took the memory and knowledge
tests only after viewing the videotapes. This design
permitted assessment, with appropriate controls, of both
within-subject and between-subject changes in knowledge,
memory, clustering, and strategy use.
Testing Sequence for Three Groups of Children Tested
for Their Memory and/or Knowledge of Birds
Memory Knowledge Memory Knowledge
Test Tests Videotapes Test Tests
Within-subject analyses for changes in knowledge
compared pretest versus posttest scores on the grouping,
listing, facts, and matching tests of the Knowledge Test Only
group. Within-subject analyses for changes in memory,
strategy use, and clustering compared the pretest versus
posttest scores of the Memory Test Twice group.
Between-subject analyses for differences in knowledge
compared the Knowledge Test Only group's pre-VlDEO grouping,
listing, facts, and matching test scores with the other two
groups' post-VIDEO scores. Finally, between-subject analyses
for differences in memory, strategy use, and clustering
compared the Memory Test Twice group's pre-VIDEO scores with
the Posttest Only group's post-VIDEO scores.
The Posttest Only group was included to insure that pre-
VIDEO testing was not biasing the results. It also was
important to include it in order to generalize the results
with confidence to other laboratory experiments. In other
experiments, subjects typically come to the laboratory with
pre-existing knowledge. Therefore, it was important to be
able to show that children's memory for birds and knowledge
of birds was not simply changing because of a practice effect
of taking the same tests twice.
Thus, there were two reasons for designing the study to
include both within- and between-subject analyses. One was
the need to include particular controls for repeated testing.
The second was to facilitate the comparison between this
study, which attempted to construct knowledge, and previous
studies, which examined pre-existing knowledge.
All children were tested individually. Children gave
their answers for all questions aloud, and the experimenters
tape-recorded and/or wrote what each child said on the data
sheets. No tests had time limits.
The memory test lasted approximately 10 minutes per
child. The knowledge tests were administered in a different
session, which always followed the memory test, if it were
given, and took approximately 35 minutes per child. Most
children were tested for both memory and knowledge on the
same day, but for practical reasons several were given the
knowledge test the day after they took the memory test.
One male undergraduate student administered all of the
memory tests, and one female graduate student administered
all of the grouping and listing tests. Finally, one of two
female undergraduates, depending upon each one's
availability, administered the facts and matching tests.
During the first week, children were tested either for
their memory for a list of birds via a recall task or for
their knowledge of birds via four different tasks. After
three additional weeks had passed, children were shown the
five videotapes, one per day, in their classroom by their
classroom teacher. Finally, post-VIDEO tests were
administered the following (sixth) week.
Children entered a small room inside an office complex
at their school. The experimenter introduced himself and
told them what they would be doing. He explained that the
children were being videotaped, so we could compare how they
played the game to how other children played the game.
Two pieces of black tagboard sandwiched the array of 20
picture cards. There were four cards per row and five cards
per column. Only the corners of the cards touched.
The experimenter lifted the tagboard, to reveal the
picture cards, and instructed children to read the name of
each bird once aloud. If a child mispronounced a bird's
name, the experimenter pronounced it correctly. Then, the
tagboard was replaced.
The experimenter instructed children to study the birds
in any way they wished for two minutes. He also told them
(and demonstrated to them) that they could move the cards
around if they wished. Half of the children studied the
cards for List A and the other half studied List B. After
two minutes of study time, the experimenter closed the top of
the tagboard, so no cards were visible. To eliminate the
recency effect, children received a sheet of math problems,
which contained simple addition problems without carrying
that they computed for 30 seconds. When the experimenter
said "stop," they were instructed to say as many birds' names
as they could remember in any order. After 15 seconds of
silence, the experimenter said, "Can you remember any other
birds?" After an additional 15 seconds of silence, the
testing session ended.
l. Grouping . Before going to a child's classroom, the
experimenter shuffled the picture cards and placed them
randomly into five rows with nine cards in each row on a
table at which the child would be seated. No more than two
cards from the same category ever appeared consecutively in
either direction. Children walked with the experimenter to
the room, which was upstairs in a loft at their school
library. While they were walking, the experimenter
introduced herself and said:
We're interested in what kids your age know about birds.
You're going to be doing four different activities. We
have been doing these activities with adults and find
that they don't know very much about birds, so don't be
upset if there are birds you've never heard of before.
When they reached the room, the experimenter explained
that the first activity was to decide which of the birds were
alike or went together in some way. She told children to
read each bird's name aloud as quickly as possible and
assisted them with names they could not pronounce.
As children gave the experimenter birds they thought
belonged together, she wrote them on a data sheet. Then,
after asking the children how they thought the birds were
alike or went together, she wrote the children's chosen title
for the group on the data sheet too. This procedure
continued until children had eight groups of birds, used all
45 birds in groups, or could not think of any other groups
for the remaining birds, whichever came first.
2. Listing . After children completed the grouping task,
the picture cards were removed from the table, and they heard
the directions for the listing task. The experimenter said
"Tell me as many birds as you can think of that are " .
One of the following six categories was read: birds of prey,
birds people eat, talking birds, birds awake at night, birds
that live by the water, and songbirds, in that order. The
experimenter wrote all the birds children said under the
appropriate category on the data sheet. Children were
prompted by the experimenter after 15 seconds of silence.
After an additional 15 seconds of silence, the experimenter
said the next category. This procedure continued for all the
3. Facts . After the children completed the listing task,
the experimenter thanked them for doing those activities and
introduced them to another female experimenter, who explained
the facts task. Children received three lists of birds
counterbalanced for order. (Each column of Appendix D was
printed on a separate piece of paper, so the order could be
changed.) They said each bird's name aloud into a tape
recorder, and then said as many facts as they knew about the
bird. If they did not know any facts about that particular
bird, they were instructed to say "don't know." The
experimenter told them not to guess, only to say facts , and
not to spend too much time thinking about a bird.
4. Matching . After children completed the facts task,
the experimenter explained that they would be matching birds
and categories for the final activity. After placing the
tagboard strip on their lap, she told the children the names
of the six categories, simultaneously pointing to each one,
in turn, on the tagboard display. The experimenter read the
names of birds from a data sheet (see Appendix E), and wrote
the numbers that represented each category, as children
responded. Children pointed to the category to which they
thought each bird best belonged. They were told to guess if
they did not know a bird.
The four knowledge tests always were given in this
order: grouping, listing, facts, and matching. The grouping
test had to be first, so that children would not be biased in
their groupings of birds by the categories they heard on the
listing and matching tests. Matching could be last, because
it was a test for knowledge that could not be acquired by
taking the other tests.
Before each classroom teacher showed her class the
videotapes, she told the children to try to remember as much
information as possible. She told them they would see five
videotapes about birds over the week, and they would be
tested the following week.
Coding and Scoring
The number of birds correctly recalled from the list of
birds was tallied. The criterion used for counting a word
was that the child had to pronounce at least one syllable
from the word correctly. For example, "whip" was counted for
"whip-poor-will." Words were counted only the first time the
child said them (i.e., no credit was given for repetitions).
1. Grouping . The labels children gave to their groups of
birds were examined. One of three possible designations was
given to each label. If the label contained information that
clearly was visible from the pictures of the birds (e.g.,
birds that are black, birds with big feet, birds with small
beaks), then that label was considered a "physical
description." If the label contained information that was
contained in the names of the birds (e.g., birds that end in
"bird," birds that begin with "s"), then that label was
considered "sound-based." If the label did not contain any
information that could be obtained from the picture of the
bird or the name of the bird (e.g., gamebirds, petbirds,
colorful birds), then the label was considered "meaning-
based." The total number of birds a child included in
meaning-based groups was tallied. Taught birds and control
birds were scored separately for analysis. Independent
scorers reached 9 5% agreement when scoring half of the
subjects' data in this manner. In the event of disagreement,
a third independent rater scored that child's data. All
disagreements were resolved in this manner.
2. Listing . The total number of birds correctly listed
under the appropriate categories was counted. Appendix F
displays the answers that were considered acceptable and
unacceptable for each category. These were determined by an
ornithologist. All categories except "talking birds" were
counted. The category "talking birds" was considered the
control set, because that category, and the birds contained
within it, was not taught by the videotapes.
3. Facts . The number of correct facts children gave for
each bird was tallied, and then the scores for all birds were
summed. Books about birds (Bull & Farrand, 1977; Peterson,
1947; Robbins, Bruun, & Zim, 1966; Steward, 1977; Zim &
Gabrielson, 1956) or an ornithologist were consulted when
questions about the truth of certain facts were encountered.
The correlation between the scorings of two independent
raters of 20 children's data was .97.
Other measures calculated were the number of birds about
which children told at least one fact, and the percentage of
the total correct facts they told that were of a general,
categorical nature, rather than specific in nature. For
example, if a child said that a bird "lived by water," "ate
animals," or "sang a song," those facts were considered
general, categorical statements. On the other hand, if a
child said facts such as a bird "lived by the ocean," "ate
mice," or "sang its name", those facts were considered
specific in nature, and therefore were not counted.
4. Matching . Appendix E displays the scoring key
utilized for the matching test. Again, separate scores were
obtained for taught birds and the control set of birds.
The Adjusted Ratio of Clustering (ARC score) was used as
an index of clustering in the children's recall. Roenker,
Thompson, and Brown (1971) reported that this measure had the
advantage of allowing comparisons of relative amounts of
clustering over trials and across experiments, because it
reflected the "proportion of actual category repetitions
above chance to the possible category repetitions above
chance." (p. 45) The formula for the ARC score is provided
in Appendix G.
Overt Study Strategies
Because it was important to assess spontaneous strategy
use in this study, only overt strategies could be assessed.
The two minute study time was videotaped and coded by two
independent observers, who reached 88% agreement on whether
each of the four categories of study behaviors occurred for a
subsample of 40 children (see Appendix H) . Four aspects of
overt study strategies were coded: involvement with the task
of remembering, rehearsal, involvement with the cards, and
self-testing. In the event of discrepancy, a third
independent rater watched and coded the videotape for that
child. All disagreements were resolved in this manner.
Behaviors designated as "strategic" are starred in
Appendix H. The presence versus absence of each type of
study behavior (coded as "l" if it had at least one checkmark
by a "strategic" behavior and "0" if it did not) was
correlated with recall. In addition, a composite strategy
score was correlated with recall. A maximum of four points
(one for each study strategy category) could be earned toward
the composite score that represented total strategy use.
This composite score also was utilized to assess whether
increases in study strategies occurred after children watched
the videotapes about birds.
The results are subdivided according to the most
important hypotheses of this study. The first main question
was whether children's knowledge of birds increased
significantly after they viewed the videotapes about birds.
In accordance with the design of the study, this question was
answered via both within-subject analyses and between-subject
analyses . The second important question was whether
children's memory, clustering, and strategy use changed
significantly after they viewed the videotapes. The current
theories about memory development would predict that if
knowledge changed significantly, then these measures also
should change significantly. Again, the answer to this
question was investigated via both within-subject and
between-subject analyses. The third and final question was
whether there were correlations among knowledge, strategy
use, and recall. This relation has been demonstrated by
previous memory development research.
Did Knowledge Change?
The only children used in the analyses for the within-
subject changes in knowledge were the Knowledge Test Twice
group (refer to Table 2-1 for the design of the experiment).
The means and standard deviations for each test are displayed
in Table 3-1.
A z-score was computed for each child on each of the
four knowledge pretests by subtracting the mean for that test
from the test score a child achieved on that test, and
dividing that difference by the standard deviation
Pre-VIDEO and Post-VIDEO Means and Standard Deviations
For Each Knowledge Test
Note: All z-scores were computed using the Pre-VIDEO means
and standard deviations.
for that test. Second, an average pretest z-score for each
child was computed by adding together the child's z-scores
for each of the four tests, and dividing the sum by four.
Third, in order to look at possible improvements in
knowledge, each child's posttest scores for each knowledge
test were converted to z-scores (according to the pretest
means and standard deviations so that change would be
apparent). Finally, average posttest z-scores for each child
were computed in the same manner as for the average pretest
To see if the videotapes caused children's knowledge to
improve significantly, the overall measures of children's
knowledge (z-scores) were compared. An Analysis of Variance
(ANOVA) examined the effects of age (2: older versus
younger), sex (2), and session (2: pretest versus posttest)
on the average z-scores. Session was a within-subject
variable. There was a significant main effect of age,
F(l,15) = 6.17, p < .05. Older children had higher z-scores
than younger children. More importantly, however, was the
fact that z-scores were significantly higher for the posttest
than they were for the pretest, F(l,15) = 19.30, p_ < .001.
Thus, as hypothesized, knowledge of birds improved
significantly after children watched the videotapes about
In order to determine whether or not all knowledge tests
changed significantly from pretest to posttest, or if some
groups improved to a greater extent than others, an ANOVA was
computed for each knowledge test separately. Each 2 (sex) by
2 (age) ANOVA had repeated measurements on session (2).
1. GROUPING Test . The ANOVA for the GROUPING test
scores revealed significant main effects of age, F(l,18) =
6.23, 2 < - 05 ' and session, F(l,18) = 15.14, p < .01.
However, there also were significant interactions between
session and age, F(l,18) = 8.05, p < .05, and session and
sex, F(l,18) = 5.69, p < .05.
Follow-up analyses were conducted to investigate these
interactions. ANOVAs were computed for each age separately.
Sex was the grouping variable in each, and session was a
repeated measurement. Whereas older children improved
significantly from pretest to posttest (means were 16.50 and
25.83, respectively), F(l,10) = 33.97, p <.001, younger
children did not (means were 8.70 and 11.10, respectively).
Other ANOVAs were computed for each sex separately. Each used
age as the grouping variable and session as a repeated
measurement. Whereas males improved significantly from the
pretest to the posttest (means were 11.67 and 21.33,
respectively), F(l,10) = 22.20, p < .001, females did not
(means were 14.50 and 16.50, respectively). Therefore, the
interactions were due primarily to older children and males
showing more improvement from pretest to posttest in
categorizing birds as members of meaning-based groups and
less tendency to categorize birds by physical appearance or
the sounds of their names.
A parallel ANOVA was conducted using the number of the
control set of birds (i.e., those not mentioned on the
videotapes) that were placed into meaning-based groups.
Unlike the results for the birds that were taught, no changes
occurred for any group in the number of control birds that
were placed into meaning-based groups. Thus, it is likely
that the increase in the number of taught birds that were
placed into meaning-based groups was due to the videotape and
not simply an artifact of repeated testing.
2. LISTING Test . The ANOVA that investigated the
influence of age (2), sex (2), and session (2) on listing
test scores also showed significant main effects of age,
F(l,18) = 5.75, p < .05, and session, F(l,18) = 84.44, p_ <
.0001. However, there also was a significant three-way
interaction between age, session, and sex. Although all
groups improved from pretest to posttest, some groups
improved more than others (see Figure 3-1). Younger males,
F(l,4) = 8.96, p < .05; younger females, F(l,4) = 72.00, p <
.01; and older males, F(l,6) = 106.67, p < .001, all were
able to list significantly more birds for each category after
watching the VIDEO about birds . Although older females were
able to list more birds after the VIDEO, they did not improve
significantly (p = .09). Thus, the three-way interaction in
the ANOVA for listing was due primarily to differential
improvement within groups .
A parallel ANOVA using only the number of talking birds
(i.e., a category about which no information was provided on
Pretest and Posttest Listing Test Scores for
Older Versus Younger Males and Females
the videotapes about birds) that children listed before and
after the videotapes confirmed that no significant
improvement occurred in this control measure for any group.
Thus, the observed changes in the taught birds probably were
not simply an artifact of testing.
3. FACTS Test . The ANOVA for the total number of facts
generated for the 45 birds had neither significant main
effects nor significant interactions. Children's ability to
list facts for birds was not influenced by the videotapes.
Likewise, the number of facts known for the control set of
birds did not change significantly.
Although the total number of facts known about birds did
not change, an analysis was conducted to determine whether
the number of birds about which children knew information
changed. Did children learn information about birds that
they did not know before the videotapes? The number of birds
about which children knew at least one fact was counted and
used in this ANOVA. This ANOVA also showed no effect of
session for any group.
Although the two previous ANOVAs suggested that the
number of known birds and the number of facts generated by
children remained constant, a final ANOVA was conducted to
determine whether the nature of the facts children generated
changed as a result of watching the videotapes. The
percentage of facts that were general, categorical statements
was utilized for this analysis.
The increase in the number of facts that referred to
categorical information after children watched the videotapes
was striking. The percentage of general, categorical
statements children said for the posttest (M = 62.05) was
significantly higher, F(l,15) = 14.70, p < .05, than the
percentage for the pretest (M = 43.16). There were no
differences among older and younger children or males and
females. Thus, the type of facts children said differed in a
more qualitative than quantitative manner.
4. MATCHING Test . The ANOVA that investigated the
effects of age, sex, and session on matching test scores
revealed a significant main effect of session, F(l,18) -
4.13, p < .05. Children were able to match a significantly
larger number of birds with their appropriate categories
after watching the videotapes about birds. The means for the
pretest and posttest matching tests were 22.09 and 24.09,
respectively. On the other hand, the control set of birds
showed no corresponding change in the number of correctly
matched birds. Again, the change appeared to be due to the
videotapes rather than an artifact of testing.
To summarize the within-subject analyses, three of the
four knowledge tests had significant improvements from the
pretest to the posttest quantitatively. Older children and
males improved their ability to group birds by meaning to a
greater degree than younger children and females. Older
females did not improve their ability to list birds under
appropriate experimenter-designated categories significantly,
but the other groups did. Both males and females improved
their matching of birds with categories. Although neither
the number of birds for which children told facts nor the
total number of facts children could generate changed
significantly, the nature of the facts children told changed
from more specific information about individual
birds to more general, categorical information. Therefore,
knowledge did improve, especially for older children and
males, and in a both a quantitative and qualitative manner.
Preliminary ANOVAs were computed to investigate whether
there were any differences among the groups on posttest
knowledge (i.e., posttest z-scores). In order to safely
assume that the pretest scores of the Knowledge Test Only
group were representative of children's knowledge of birds,
and to generalize the results to standard laboratory tasks
(in which children come to the memory task with pre-existing
knowledge), it was important to demonstrate that the groups'
posttest knowledge scores did not differ. Fortunately, the
results confirmed that there were no significant differences
among the groups. Therefore, for all between-subject
assessments for post-VIDEO knowledge, the Recall Test Twice
and the Posttest Only groups' scores were combined. These
post-VIDEO knowledge test scores were compared to the pretest
(or pre-VIDEO) scores of the Knowledge Test Only group (refer
to Table 2-1 for the design of the study) to assess
hypothesized differences in knowledge.
A Multivariate Analysis of Variance (MANOVA) was
conducted, looking at the effect of age, sex, and time of
testing (2: pre-VIDEO or post-VlDEO, depending upon the
group) on the scores for each knowledge test. MANOVA was
considered the best analysis, because it takes the
correlations among the tests into consideration. There was a
significant main effect of age, F(4,80) = 2.77, p < .05.
Overall, older children had higher knowledge scores than
younger children. According to the Pillai's Trace statistic,
there also was a significant main effect of time of testing,
F(4,80) = 4.53, p < .01. Post-VIDEO scores were higher than
pre-VIDEO scores. Thus, these results confirm those of the
previously presented within-subject analyses.
In order to determine which of the knowledge tests
demonstrated significant improvement, ANOVAs were computed
for each knowledge test separately. Every 2 (age) by 2 (sex)
by 2 (time of testing) ANOVA showed main effects of age:
F(l,83) = 8.59, p < .05, for grouping; F(l,83) = 4.95, p <
.05, for facts; F(l,83), p < .01, for listing, and F(l,83) =
4.91, p < .05, for matching. Older children's scores were
superior to younger children's scores on every knowledge
test. The ANOVA for matching also showed a significant main
effect of session, F(l,83) = 5.29, p < .05. Children were
able to match a significantly higher number of birds with
their correct categories after they saw the videotapes (pre-
VIDEO versus post-VIDEO means were 22.09 versus 24.26,
respectively). The ANOVA for grouping revealed a significant
interaction between session and sex, F(l,83) = 4.22, p < .05.
Follow-up analyses of this interaction supported the results
of the within-subject analyses for grouping. Whereas boys'
ability to group birds into meaning-based groups improved
significantly, girls' ability did not. Listing did not
improve significantly after the videotapes.
The three between-subject ANOVAs investigating the three
different measures for the facts test replicated the within-
subject analyses precisely. There were no significant
differences between pre-VIDEO and post-VIDEO scores
representing the total number of facts generated by children
or the number of birds for which children knew at least one
fact. On the other hand, the percentage of facts that were
general, categorical statements was significantly higher for
the post-VIDEO (M = 57.94) than the pre-VIDEO (M - 43.16)
test. Thus, grouping and matching knowledge improved from
pre-VIDEO to post-VIDEO testing, especially for boys, and
children's fact generation changed in a gualitative manner.
Did Recall, Clustering, and Strategy Use Change?
Analyses for within-subject changes in recall,
clustering, and strategy use compared the pretest and
posttest scores of the Memory Test Twice group (see Table 2-1
for the design of the study). Current theories of memory
development would predict that if children's knowledge of
birds improved significantly, then recall, clustering, and
strategy use should also improve.
l. Memory Test . Preliminary analyses showed no
significant differences between recall of List A or List B.
Therefore, children who recalled both lists were combined for
subsequent analyses. An ANOVA was computed to assess the
effects of age (2), sex (2), and session (2: pretest versus
posttest as a within-subjects variable) on the memory test
scores. There was a significant main effect of sex, F(l,39)
= 7.16, p < .05, with boys recalling more words than girls,
overall. There also was a significant sex by session
interaction, F(l,39) = 10.99, p < .01 (see Table 3-2).
Follow-up ANOVAs looked at the effect of age and sex on
recall for each session separately. Whereas boys recalled a
significantly higher number of birds than girls on the
pretest, F(l,39) = 19.07, p < .001, there were no differences
between boys' and girls' recall on the posttest. Additional
follow-up ANOVAs looked at the effects of age and session on
recall separately for boys and girls. Despite the fact that
Means (and Standard Deviations) of Pretest and Posttest
Recall for Males and Females in the
Memory Test Twice Group
Sex n Pretest Recall Posttest Recall
only boys gained significantly more knowledge of birds on the
grouping test after the VIDEO, and boys' improvement was, at
minimum, equal to girls' improvement, it was only the girls
who improved significantly on the memory test. This finding
was contrary to what would have been predicted by current
memory development theories .
2. Clustering . It would not make sense to consider
change from below chance clustering (ARC scores less than
zero) to chance clustering (ARC scores equal to zero) as
improvement. Therefore, for the purpose of analyses of
changes in clustering, ARC scores less than or equal to zero
were all set to equal zero. A score of zero meant that no
identifiable clustering was present in the child's recall.
An ANOVA was conducted to investigate the influence of age
(2), sex (2), and session (2) on the ARC scores. Although
current research debates whether clustering reflects a
deliberate strategy or automatically activated associations,
the fact that intra-group associations were emphasized in the
videotapes would lead one to predict that clustering would
improve after the videotapes, whether or not children were
aware of clustering as a strategy. However, this was not the
case. There was neither a significant main effect of
session nor any significant interactions involving session.
3. Strategy Use . An ANOVA was computed to see if there
were any effects of age, sex, and session on children's
composite strategy scores (i.e., involvement with the task,
rehearsal, involvement with the cards, and self-testing).
Appendix H provides the behavioral descriptions for the
strategy categories which comprised the composite score.
Again, current research would predict such changes when
children's knowledge increased. However, the results showed
no significant main effects and no significant interactions.
Therefore, like clustering, strategy use did not change for
As was the case for the knowledge tests, preliminary
analyses showed no significant differences among the groups
on posttest memory test scores. The only two groups compared
were the Memory Test Twice group and the Posttest Only group,
because the Knowledge Test Only group did not take the memory
test (refer to Table 2-1 for the design of the study).
Therefore, for the between-subject analyses, the pre-VIDEO
scores of the Memory Test Twice group were compared with the
post-VIDEO scores of the Posttest Only group.
1. Memory Test . Preliminary analyses revealed no
significant difference between the memory test scores for
children who studied List A versus those who studied List B.
Therefore, further analyses did not include list as a
An ANOVA examined the effects of age, sex, and time of
testing (pre-VIDEO versus post-VIDEO) on the memory test
scores. There were neither significant main effects nor
interactions. These results again were contrary to what
would have been predicted by current memory development
Means (and Standard Deviations) of Pre-VIDEO and Post-VIDEO
Clustering for Younger and Older Children
Age Clustering Clustering
Younger .29 .06
( .36) ( .10
Older .21 .27
( .25) ( .30
2. Clustering . The ANOVA that investigated the effects
of age, sex, and time of testing on clustering revealed only
a significant age by time of testing interaction, F(l,67) =
4.26, p < .05 (see Table 3-3). Contrary to the expected
improvement in clustering predicted by current theory, the
interaction was due to the significant decreases in
clustering by younger children after the videotapes, F(l,30)
= 4.52, 2 < .05. The older children's clustering remained
unchanged by the videotapes.
3. Strategy Use . Supporting the within-subject analyses
of strategy use, the ANOVA that investigated the influence of
age, sex, and time of testing on children's overt strategy
use showed neither significant main effects nor significant
interactions. This, too, was contrary to what would have
been expected by current theory.
Individual-Subject Analysis of Changes
in Knowledge and Memory
Because group means do not always accurately reflect the
performance of individuals, an individual-subject analysis of
changes from the pretest to the posttest for all knowledge
tests and the memory test was conducted. This analysis was
expected to reveal whether the significant changes in pretest
to posttest scores were due to a substantial number of
children's test scores improving, or if they were due only to
a few children who improved substantially. An individual-
subjects analysis also was expected to reveal whether the
lack of significance for the memory test was due to few
subjects improving, or whether most subjects improved but
only by a few points. Table 3-4 reveals that at least 50% of
the younger children's scores improved for every knowledge
Individual-Subject Analysis for Patterns of Changes
on the Knowledge and Memory Tests by Age
Direction of AGE
Change from Younger 3 Older'
Pretest to Posttest
Facts Test: Total
Facts Test: Number
Facts Test: Type
Memory Test 43 (2.10) 33 (2.00)
Grouping Test 40
Listing Test 10
Facts Test: Total
Facts Test: Number 20
Facts Test: Type 11
Matching Test 10 17
Memory Test 24 3 3
Grouping Test 40 (4.25)
Listing Test 8 (3.00)
Facts Test: Total 40 (11.25) 58 (5.86)
Facts Test: Number 20 (3.50) 22 (6.00)
Facts Test: Type 11 (5.20)
Matching Test 40 (1.25) 25 (2.00)
Memory Test 33 (2.90) 33 (1.70)
Note: Numbers reflect the percentage of children at each age
whose pretest to posttest scores were in the stated
direction. The average improvement or regression for
the group is in parentheses .
a n = 10 for the knowledge tests and n = 21 for the memory
" n = 12 for the grouping, listing, and matching tests, n = <
for the facts test, and n - 21 for the memory
test measure except grouping, and at least 50% of the older
children's scores improved for every knowledge test measure
except the total facts generated. Especially noteworthy were
the improvements by 100% of the younger children (an average
of 16.79 percent) and 78% of the older children (an average
of 27.01%) on the percentage of generated facts that referred
to general, categorical information.
In contrast to the knowledge tests, the pattern for the
memory test appeared to be due more to chance variation than
to consistent but small improvements by the majority of
children. There were approximately two points difference
between pretest and posttest scores on the average for both
children showing improvement and children showing regression
on the memory test.
Were There Significant Correlations Among Knowledge,
Clustering, Strategy Use, and Memory ?
Correlations were computed between each of the knowledge
test scores, clustering (ARC scores), or composite strategy
use scores, and the memory test scores for younger children,
older children, and all children (see Table 3-5). Although
the correlations between each knowledge test and recall were
significant for the sample as a whole, the correlations were
higher for older than for younger children. Multiple
regression analyses were computed for post-VlDEO memory test
scores as a function of the four knowledge tests for each age
Post-VIDEO Correlations Between Knowledge, Clustering,
or Strategy Use and Recall for Groups of Children
separately. Whereas the four knowledge tests explained 26%
of the variation in older children's memory test scores,
they only explained 6% of the variation in younger children's
memory test scores. Therefore, knowledge of an area becomes
a better predictor of memory with age.
As Table 3-6 shows, the correlations among the four
knowledge tests all were higher for older children, with the
exception of those between listing or matching and facts
(which were each .02 higher for younger children). The
largest discrepancy between older and younger children's
correlations was between the listing and facts tests. Thus,
older children's knowledge is more inter-related than younger
Correlations Among the Knowledge Tests By Age
To determine whether clustering predicted recall, two
Analyses of Covariance (ANCOVA) were computed. Each
investigated whether clustering predicted recall, while
controlling for differences in age and sex. The results for
the ANCOVA that investigated pre-VlDEO clustering were that
there was a significant effect of clustering, F(l,38) = 4.60,
p < .05. Children who had more clustering in their recall
remembered more words. The three variables in this analysis
explained 38% of the variation in recall scores. It is
interesting to note that sex was still significant (as
discussed in the within-subject follow-up tests for the sex
by trial interaction on the recall measure), even when age
and clustering were controlled. However, there were no
significant age differences in recall. The results of the
ANCOVA that investigated post-VlDEO recall also had a
significant effect of clustering, F(l,71) = 10.24, p < .01,
controlling for differences in age and sex. Again, higher
clustering was associated with better recall, and the three
variables explained 17% of the variation in recall scores.
The results showed no effects of sex or age. Therefore,
clustering did predict recall, as one would expect.
Summary of the Results
The hypothesis that knowledge would increase from the
pretest to the posttest was supported by both within- and
between-subject analyses. This increase, however, did not
occur in all subject groups on all assessments of knowledge.
All children significantly improved their ability to match
birds with their correct categories after seeing the
videotape about birds. However, for the grouping test, the
change was significant only for boys. That is, girls did not
significantly improve their ability to place birds into
meaning-based groups. In addition, older girls did not
significantly improve the number of birds they could list for
The only knowledge test that did not show improvement
quantitatively was the facts test. Two different measures
(i.e., the total number of facts children could generate for
the birds and the number of birds for which children knew at
least one fact) did not improve after children saw the
videotapes. In fact, the post-VlDEO means were slightly
lower than the pre-VIDEO means. However, the nature of the
generated facts changed. Whereas few responses for birds
contained general, category information during the pre-VIDEO
test, a significantly higher percentage of the facts said
during the post-VIDEO test contained that type of
information. This change was evident for both within- and
between-subject analyses. Because the pictures of the birds
were not visible for this test, as they had been in the Chi
and Koeske (1983) study, it was unnecessary to distinguish
between facts directly related to the picture and facts not
observable from the picture.
With respect to the second hypothesis, despite
significant increases is knowledge, there were no
corresponding improvements in memory, clustering, or strategy
use. Although there were significant differences in knowledge
between older and younger children, there were no age
differences in recall. Boys did not improve their recall
after seeing the videotape about birds, even though they
clearly gained more knowledge about birds than girls. Girls
also did not improve their recall according to the between-
subject analysis but did on the within-subject analysis.
Because there was so much fluctuation (i.e., both
progressions and regressions) on the memory test, according
to the individual-subject analysis, there is not a lot of
evidence to indicate positive change. Only two younger
females and two older females had recall scores that changed
more than two points. Clustering and strategy use did not
improve for any group. Therefore, knowledge improvements did
not translate into better strategy use or memory, contrary to
the predictions of current theories of memory development.
This suggests that improvements in knowledge either are not
necessary for improvements in memory with age, or that
knowledge is necessary, but not sufficient, for the
With respect to the third hypothesis, consistent with
previous studies which were correlational in nature, each
knowledge test had a positive correlation with the memory
test. Older children's knowledge and memory had slightly
higher correlations than did younger children's.
The results of the present study are important because
they raise certain questions that might lead to a clearer
specification of the relation between knowledge and memory.
Previous research that credits knowledge as being responsible
for the major changes in memory with age has been criticized
for using the term knowledge too globally and ambiguously
(Saarnio, 1987). In contrast, this study included four
different assessments of knowledge and thus could provide a
more refined description of knowledge. Each of the four-
knowledge tests presented a slightly different picture of the
relation between knowledge and memory. Thus, all assessments
of knowledge are not equal, and it is important for future
studies to include several measures of knowledge.
The fact that the correlations between knowledge and
memory increased with age supports the suggestion from
previous research that categorical information is utilized
for mnemonic activities to a greater extent by older children
than younger children. The younger children appeared to be
more influenced by salient names or perceptual features. For
example, practically every third grader recalled the birds
lovebird and titmouse, perhaps because of salient
associations. Yet, these same children knew nothing about
those birds, and they tended to group them together with the
cardinal and bluejay, labeling the group "birds with things
on top of their heads," a characteristic clearly visible from
the pictures of those birds. Although very few older
children knew information about those same birds, they also
rarely recalled them.
In conclusion, when the present study looked at the
relation between knowledge and memory at one point in time,
it replicated the previously documented correlation between
knowledge and recall. However, a different picture emerged
for the group comparisons in the present study that examined
the effect on recall of constructing new knowledge, a
comparison that was not possible in previous studies because
they did not include this manipulation. Improvement in
knowledge did not lead to increased recall. That new
associations between birds and their category membership were
created was apparent from the increases in post-VlDEO
knowledge test scores. Yet, like many studies showing that
younger children "have" the available knowledge but do not
utilize it, a common finding in research on the production
deficiency, the children in this study apparently did not use
the new knowledge to enhance their study strategies,
clustering, or recall. Thus, this study suggested that
acquiring new knowledge may not guarantee its access.
Given evidence for correlation but not causality, the
next question logically becomes "Why?". Interpretations of
the data will be discussed in the next section.
Why Correlation But Not Causality?
There are two main explanations for why increases in
knowledge did not lead to increases in recall. One has to do
with the restructuring of knowledge and the other involves
speed of activation of related information.
One possible explanation for why the new knowledge was
not accessed is that although the children acquired the
associations between the birds and their categories, they did
not restructure that knowledge into a usable form. Chi
argues that having categorical associations is not the same
as having a semantic structure (Chi & Ceci, 1987). Although
intra-group associations were taught in the videotapes and
learned by children in the present study, the children may
not have acquired a class inclusion relation (i.e., that all
birds in a category simultaneously have certain
characteristics and are members of the category and, at the
same time, are part of a larger category, birds). As a
result, they might not have had access to all the birds
within a category simultaneously. In this case, children may
not spontaneously generate category labels and use them to
generate category members. Therefore, this explanation for
the results would claim that knowledge may need to be in a
particular form to become accessible. In other words,
quality of knowledge, as well as quantity, is important.
With further exposure to information about birds, children
might have reorganized their knowledge so that it would have
become more accessible to them.
More specifically, knowledge could be structured to be
more accessible in the following way. Knowledge may
initially be represented in linear associations, next in
multiple associations, and finally in hierarchically
organized associations. If any associations initially were
simply linear, then activating any member of the pair would
only activate the other member. Eventually, interconnections
among items would increase so that knowledge would be
represented as a network. Take, for example, the analogy of
a spider web, where each item may be connected to many other
items, but the network itself is fragile. Activating any
part of the structure would make accessing any other part of
the structure more likely, but the whole structure still
wouLd not be interconnected. At this point, children are
unlikely to generate a category name and use it as a cue for
retrieving birds. Eventually, as a greater and greater
number of associations between birds are established, and
previous associations become stronger, the form of knowledge
might change to become a hierarchical structure. The members
would be perceived as a unit--one of several units of the
bird kingdom. Activation of any part of the structure would
simultaneously activate all the other parts; recalling any
single bird should activate the entire structure and lead to
the accessing of all birds in that category. Knowledge is
easily accessed, for one need only retrieve the whole
Thus, according to this position, memory development
would be the result of both qualitative and quantitative
changes in the structure of knowledge. These structural
changes would make knowledge more accessible. Using the
above analogy, perhaps knowledge was changed to a form
similar to the spider web for many children in the present
study. They were able to list more birds for categories and
could group a larger number of birds with self-titled
meaning-based categories. They generated a higher percentage
of general, categorical facts after watching the videotapes.
Thus, they had more than single, linear associations.
However, they may not have been able to retrieve the whole
A second possible explanation is that the created
knowledge was not accessed because the exposure to the
videotapes did not increase the speed at which the new
knowledge could be activated, perhaps because of the way the
information was taught. During the videotapes, associations
between birds were taught. In an effort to equate the
teaching of all birds, no more than two birds from the same
category ever were compared with any other bird in that
category. Therefore, all the new associations received equal
attention, and probably had equal strength.
Perhaps for knowledge to be utilized, some pathways need
to be traveled more frequently, thus creating stronger routes
to knowledge rather than simply additional pathways to that
knowledge. Bjorklund (1987) claims that increases in the
frequency of occurrence of items probably cause an increase
in the number and type of features and create more links with
other items in semantic memory. However, the results of this
study suggest that simply creating more links (by teaching
facts about birds) does not appear to be sufficient, and
although the number of links from categories to birds or
birds to categories increased in this study, that too was not
sufficient for improved memory. Perhaps frequent exposure to
items is necessary so the links between items become well-
practiced, and it is only after the information can be
activated more readily that knowledge is utilized. More
specifically, repeated exposure to swans, geese, and ducks in
the context of water may result in water and the names of
these birds immediately becoming activated when any single
bird is recalled or encountered.
The correct answer to "What is memory the development
of?" might be "knowledge, ease of access, and strategies, in
that order of acquisition." This answer would suggest that
knowledge and ease of accessing that knowledge are precursors
for strategy development, and that increased strategy usage,
particularly strategies that tie the material to be learned
with prior knowledge, is what increases memory performance.
More specifically, memory development might proceed as
follows: First, perhaps some knowledge of the area is
acquired. This foundation of knowledge might be necessary as
a first step. In this study, children appeared to acquire
categorical knowledge mainly about those birds for which they
had previous information. Evidence is that the number of
birds for which they generated facts did not change but the
type of facts they generated changed significantly (i.e.,
referred to more general, categorical information). Second,
repeated experience with the knowledge may strengthen
semantic grouping connections and weaken other facts , such as
perceptually-based facts. The outcome is a categorical
organization with knowledge structured in a more usable form
and/or with faster speed of activation of related
information. Third, easier access or faster activation means
that more sophisticated study strategies can be utilized,
because semantic relations are easily accessed. With few
exceptions, the body of research in cognitive psychology
confirms the superiority of performance on free-recall tasks
when the information is encoded semantically .
Either of the above explanations (i.e., structural or
speed changes), or perhaps some process unexplored here,
taken as intermediate steps between knowledge and memory,
would explain why having the knowledge is not sufficient for
improvements in strategy use or recall. Early research
generated by the hypothesis that younger children had a
production deficiency claimed that younger children had the
relevant categorical knowledge that could help them recall
but tended not to use it. The claim here is that one reason
for a production deficiency is that knowledge that cannot be
accessed readily due to its primitive structure or its slow
speed of activation will not be utilized. These accounts
also explain why preschool children only use strategies when
involved with very familiar material.
In other work on strategies (e.g., Gobbo & Chi, 1986;
Ornstein & Naus, as cited by Zembar & Naus, 1985a), those who
already can access knowledge readily (experts) are compared
with those who cannot (novices). Studies by Bjorklund and
others (Bjorklund & Thompson, 19 8 3; Buchanan & Bjorklund,
1988) compare very familiar knowledge (typical items) with
moderately familiar knowledge (atypical items). Studies by
Lindberg (1980) and Zembar and Naus (1985b) compare knowledge
that is very familiar (or very salient) for one group with
knowledge that is unfamiliar for that group. All of these
studies find increased use of strategies associated with the
former type of knowledge in the contrasted pair. However, it
can be argued that it is not simply a difference in knowledge
per se, but the likely differences in either the structure of
knowledge, the speed of activation of related knowledge,
both, or a factor unexplored here, that actually contributed
to the observed increased use of strategies.
In addition, it should be noted that Bjorklund's
comparisons of typical and atypical items contrast two types
of knowledge that both, for the most part, are at the upper
end of the continuum. Most people do not name watermelon as
the first response when asked to list pieces of fruit, but
readily would agree that it is a member of that category, and
they might name it readily during certain months of the year.
Therefore, although subjects' knowledge differs with regard
to the structure of categories or the speed of activation of
their atypical members, minimal training would be necessary
to facilitate the subjects' ability to access it. That would
not be true in the present study, which examines moderate
levels of knowledge.
It might be argued that including a larger number of
subjects or identifying and removing artifacts of measurement
might reveal a significant change in recall. If the lack of
improvement in recall were due only to too few subjects, then
the knowledge tests should have been influenced as well, in
fact, a greater number of subjects took the memory test twice
than the knowledge tests twice, so the statistical test for
changes in memory actually had more power. Furthermore, the
individual-subject data analysis revealed entirely different
patterns of results for the knowledge and memory tests .
Whereas the knowledge tests clearly showed that a majority of
the subjects improved (especially older children, almost all
of whom improved somewhat), the recall test showed that
almost equal numbers of children had scores that increased,
decreased, or stayed the same. Furthermore, the majority of
children's pretest and posttest recall scores were within one
point of each other. Finally, the lack of improvement in the
recall of males was not due to a ceiling effect. They
recalled only an average of 9.86 of the 20 possible words on
the pretest. Therefore, they had adequate room for
In an effort to generalize the results with certainty to
those of correlational studies, a design was used in which no
child took all the tests. However, that prevented the
within-subject measurements of both knowledge and memory for
the same children. The fact that the three conditions did
not differ on any test suggests that children can take all
the tests in future research without introducing practice
effects or sacrificing the generality of the results.
Other Issues and Future Research
In this first experimental (not simply correlational)
investigation of the relation between knowledge and memory,
care was taken not to present knowledge in such a way that
any changes in memory would appear to be trivial. A main
consideration for designing the instruction about birds was
that it was desirable to focus on two changes in specific
aspects of knowledge identified as important for memory in
previous research. One was creating a larger number of
intra-group associations, which was the major characteristic
that separated the well-known and less-familiar dinosaurs
known by a boy in the study by Chi & Koeske (1983). The
other was changing children's focus upon physical
characteristics to a focus on meaning-based characteristics,
labeled explicit and implicit propositions by Gobbo & Chi
(1986), and found to distinguish experts from novices. This
previous research suggested that these changes in knowledge
would be sufficient for changes in memory. The results of
this study indicate that these knowledge differences alone
may not be the cause of improvements in memory. Thus, future
research should attempt to specify the nature of the relation
between knowledge and memory more clearly.
In particular, future research should investigate the
possibilities that restructuring of knowledge or changes in
the speed of activation are intermediate steps to improvement
in memory. The structure of knowledge purposefully was not
taught directly. That is, the videotapes did not present a
category name and then list all of the birds in that
category. Children neither saw all the birds belonging to a
category at once nor were given the labels for all the
different categories at one time. Therefore, any changes in
memory would not be trivial. As in real life, children had
to do some construction of any hierarchical knowledge about
birds and categories. The fact that increased knowledge did
not bring increased recall suggests that the necessity to
construct the categorical knowledge was too effortful for
children to do for birds about which they had little prior
information or experience. It would be interesting to
determine whether children would learn new birds more readily
if they were provided with the hierarchy of categories from
the outset. Giving children that structure before they saw
the videotapes might facilitate their acquisition of a
knowledge structure, even if the birds were taught randomly,
as they were in this study.
One important aspect of knowledge acquisition may be
that members of a category be encountered in the same context
or at about the same time. To determine whether proximity of
learning is an important cue for restructuring knowledge, the
lessons about birds could be rearranged for one group of
children to see if that facilitated their later use of that
knowledge for recall. The lessons about all the birds from
one category would appear consecutively, instead of being
interspersed with birds from other categories. In the real
world, learning generally occurs either in the same physical
context or within a time context. For example, children
encounter swans, ducks, and geese at the lake (physical
context), or children may hear a whip-poor-will, a chuck-
will ' s-widow, or an owl at about the same time on different
nights (a time context). The children still would need to do
some construction of a knowledge structure, because the birds
would not be explicitly identified as belonging to a
category, and the nature of that category would not be
explained. Research also could examine the argument that
speed of access of knowledge is what is important. Perhaps
training children on half the number of birds, thus spending
twice the amount of time on each, would facilitate recall.
Another approach for future research is to use other
tasks. For example, a task similar to the clue game (the
child and experimenter take turns giving properties and
identifying the "creature") utilized by Chi and Koeske (1983)
might be attempted with a small number of children both
before and after training. Assessment of the structure of
children's knowledge might include a traditional Piagetian
class-inclusion task. Then the probabilities of trained
versus untrained birds being represented hierarchically could
Because an important part of this initial study was to
look at children's spontaneous strategy use, it was decided
to record spontaneous overt strategies rather than to elicit
particular strategies from children. Like other studies that
infer strategy use from overt behaviors, precision of
measurement often is sacrificed. For example, because very
few children rehearsed aloud, the definition of rehearsal
also included obvious lip movement, obvious head-bobbing, or
any combination of the two. Therefore, some children could
have been rehearsing covertly, but were not classified as
such because they produced no observable indicators.
Spontaneous strategy use actually was quite low, so the next
step would be to instruct children to use strategies. With
this design, increases in knowledge may lead to increased
ability to profit from this instruction.
One example of elicited strategies from children would
be having children sort the birds for recall. A comparison
of the well-trained versus untrained birds would be made.
Would children be more likely to group the well-trained birds
by meaning, and consequently to recall those particular birds
Another series of studies could investigate whether
changes in the speed of activation alone were necessary for
improved strategy use and memory. Children would receive
practice identifying the appropriate category for half of the
items on a list of birds. The pretest and posttest memory
test would include a measurement of the style and content of
the children's rehearsal (which they would be asked to
perform aloud). Comparisons would be made of the trained
versus untrained items. Would children be more likely to
include the trained than the untrained items in their
rehearsal sets? Would their rehearsal become more active and
contain a greater number of items from the same category when
they could access the items and category information more
The results of the present study challenge current views
of the relation between knowledge and memory development.
The question of "What is memory the development of?" is far
from answered, and it is time to rethink the old answers,
redefine possible mechanisms for change, clarify old
theories, or invent and test new ones. In particular, it is
crucial that researchers attempt to a) define and measure
knowledge with multiple tests and b) conduct more
experimental tests in which new knowledge is created. Both
are virtues of the present study.
LISTS OF BIRDS USED FOR THE MEMORY TEST
Birds of Prey
Birds People Eat
Birds of Prey
Birds People Eat
B irds Awake at Night
Birds that Live by Water
Birds Awake at Night
chuck-will ' s-widow
Birds that Live by Water
ORDER OF LESSONS PRESENTED ON THE VIDEOTAPES ABOUT BIRDS
Day 1 (17 minutes)
Introduction to Water Birds
Introduction to Songbirds
Introduction to Birds People Eat
Day 2 (12 1/2 minutes)
Introduction to Birds Awake
Great Blue Heron
Chuck-will ' s-widow
Bantam Rooster (chicken)
Introduction to Birds of
Red-tailed Hawk (Buzzard)
SAMPLE LESSONS FROM THE VIDEOTAPES ABOUT BIRDS
Like the sandpiper and the plover, a tern lives near the
beach. In fact, it only lives near salt water, never near
lakes or rivers like some other water birds.
Unlike the sandpipers, terns spend most of the daylight
hours in the air flying, and they're clumsy on land. You
won't find terns running up and down the beach the way
sandpipers do. Tern's long tails help them to make quick
turns to the right or left when they are flying.
The Royal Tern is a large tern with a heavy orange-red
bill, black cap, pale gray back and wings, white forehead and
black crest. Their tail has a slight fork (or V) in it. The
Royal Tern's wings are long and narrow, like the plover's
When terns search for food, they fly over the water with
their bill pointed downward. Then they dive into the water
from the air to catch fish. Royal Terns eat only fish.
There's usually not much material for a nest at the
seashore. Few plants can grow in loose sand that's
constantly stirred and shifted by wind. So, many kinds of
seashore-dwelling birds (such as the Royal Tern) either lay
their eggs right on the sand or else scoop out a slight hole
(like the sandpiper does) and lay their eggs in that. The
eggs of many seashore birds are splotched and streaked with
color that makes them look like a bunch of pebbles. Thus,
although the eggs are right out in the open, they're well
hidden by blending in with the beach. So, like plovers,
terns have a clever way to protect their young!
Unlike the noisy, screaming, poor songs of the bluejay,
the goldfinch sings a bright song "per-chick-o-ree. " As it
flies, it also sings a song in time with the beat of its
wings. That song sounds like "potato-chips."
The male and female bluebirds and bluejays are similar.
Not so for the American Goldfinch. In fact, the American
Goldfinch even looks different in the summer and winter.
In the spring and summer, the male is bright yellow with
white edges on black wings and tail and a black forehead. In
the winter the male looks more like the female: dull yellow,
brown with wing bars and no black on their heads.
Goldfinches are much smaller than bluebirds or bluejays.
You can recognize a Goldfinch when it's flying, because
it flies like a roller coaster up and down and up and
The goldfinch's main food is seeds. It especially likes
tiny seeds, so it has a little pointed bill to help it get
weed seeds, small grain, and some wild fruit. Because weed
seeds aren't available until late summer, baby Goldfinches
hatch much later than other birds like the bluebird and
bluejay. Goldfinches wait to lay eggs so they will have lots
of food for their babies.
The goldfinch is often called the "wild canary", because
of its yellow color and clear song.
The nighthawk 's name tells you one time of day that you
will find it flying. Can you guess when that is?
Yes, like the chuck-will' s-widow and the whip-poor-will,
nighthawks are birds that can be heard during the night. It
repeats a buzzy "beans, beans, beans." The nighthawk is
different from the chuck-will' s-widow and the whip-poor-will
though, because it often flies in the sunlight, too.
The nighthawk is not a hawk. It's actually a close
relative of the whip-poor-will and is the same size, too.
You can tell the nighthawk and whip-poor-will apart, because
the nighthawk is darker and has a longer tail that has a fork
(or V) at the end. It has a white patch on its wings, too.
It also doesn't spend most of its time on the ground like the
whip-poor-will and chuck-will ' s-widow. Instead, it is
constantly in the air, flying in a zig-zag pattern, circling,
diving, and banking.
It catches flying insects on its wing like the whip-
poor-will does. It also flies with its short-billed mouth
wide open. In one single day, one nighthawk ate over 500
mosquitoes and another one ate 2175 flying ants! The next
time you're out at night getting bitten by mosquitoes, try to
find a nighthawk to get rid of those pests!
LISTS OF BIRDS FOR THE FACTS TEST
chuck-will ' s-widow
blue j ay
SCORING FOR THE MATCHING TEST
1 = Birds of Prey
2 = Birds People Eat
3 = Talking Birds
4 = Birds Awake at Night
5 = Birds that Live by Water
6 = Songbirds
chuck-will ' s-widow
4 or 6
blue j ay
2 or 5
4 or 6
SCORING OF THE CATEGORY LISTING TEST
1) BIRDS OF PREY OR BIRDS THAT EAT ANIMALS
Acceptable Answers: hawk, vulture, eagle, kestrel,
osprey, kite, buzzard, snakebird, gull,
oystercatcher , heron, tern, owl
Unacceptable Answers: nighthawk, titmouse, guail,
bluejay, sandpiper, duck
2) BIRDS THAT PEOPLE EAT
Acceptable Answers: goose, pheasant, chicken, quail,
turkey, grouse, dove, duck
Unacceptable Answers: sparrow, titmouse, wren, any other
3) TALKING BIRDS
Acceptable Answers: cockatoo, monk, mynah, lory, parrot,
lovebird, parakeet, macaw
Unacceptable Answers: whip-poor-will, chuck-will ' s-
widow, chickadee, mockingbird, cardinal, canary
4) BIRDS THAT ARE AWAKE AT NIGHT
Acceptable Answers: chuck-will ' s-widow, whip-poor-will,
mockingbird, nightingale, owl, nighthawk
Unacceptable Answers: starling, chickadee, osprey, bat
5) BIRDS THAT LIVE BY THE WATER
Acceptable Answers: plover, oystercatcher, goose, heron,
tern, swan, snakebird, gull, sandpiper, stork,
crane, coot, flamingo, egret, duck, pelican
Unacceptable Answers: turkey, quail, swallow
Acceptable Answers: bluebird, blackbird, cardinal,
titmouse, chickadee, mockingbird, goldfinch,
sparrow, bluejay, nightingale, starling, swallow,
Unacceptable Answers: whip-poor-will, chuck-will ' s-
widow, lovebird, parrot, quail, crow, owl, lory,
dove, hummingbird, parakeet, killdeer, pigeon
FORMULA FOR THE ADJUSTED RATIO OF CLUSTERING (ARC) MEASURE
IN ROENKER, THOMPSON, AND BROWN (1971)
ARC = R - E (R)
max R - E (R)
R ■ total number of observed category repetitions (i.e.,
the number of times a category item follows an item
from the same category),
max R = maximum possible number of category
E (R) = expected (chance) number of category
It should be noted that
max R = N - k,
where N = total number of items recalled,
and k = number of categories represented in the recall
And, (from Bousefield & Bousefield, 1966, as cited by
Roenker, Thompson, & Brown, 1971)
E (R) = Sigma n^ 2 - 1
n^ 2 = number of items recalled from category i, and N is
as before. (p. 46)
STUDY STRATEGY CATEGORIES AND THE BEHAVIORAL DEFINITIONS
Involvement With the Task
* leaned forward
* stood up
* lips moving
* bobbed head up and down (at least two times
* said the names of birds aloud
Involvement With the Cards
pointed at cards (no contact made)
<5 * 5 - 10 *>10
touched the cards (contact made but the positions of
<5 * 5 - 10 *>10
moved the cards (contact made and positions of cards
*<5 *5-10 *>10
* looked up > 1 second in duration
* looked to the side > 1 second in duration
* closed eyes > 1 second in duration
* covered the names of birds with hand
Note: Starred items were considered "strategic." Children
received one point for each of the four strategy
categories that had at least one starred item checked or
circled. Then the points were summed to obtain a
composite strategy score (maximum =4).
Ackerman, B.P. (1987). Descriptions: A model for nonstrategic
memory development. In H.W Reese (Ed.), Advances in
child development and behavior (Vol. 20, pp. 143-183).
Orlando, FL: Academic Press.
Baker-Ward, L., Ornstein, P. A., & Holden, D.J. (1984). The
expression of memorization in early childhood. Journal
of Experimental Child Psychology , 37 , 555-575.
Bjorklund, D.F. (1985). The role of conceptual knowledge in
the development of organization in children's memory. In
C.J. Brainerd & M. Pressley (Eds.), Basic processes in
memory development : Progress in cognitive development
research (pp. 103-142). New York: Springer-Verlag.
Bjorklund, D.F. (1987). How age changes in knowledge base
contribute to the development of children's memory: An
interpretive review. Developmental Review , 7, 93-130.
Bjorklund, D.F. (1988). Acquiring a mnemonic: Age and
category knowledge effects. Journal of Experimental
Child Psychology , 45, 71-87.
Bjorklund, D.F. & Muir, J. (in press). Children's development
of free recall: Remembering on their own. In R. Vasta
(Ed.), Annals of child development (Vol. 5). Greenwich,
CT: JAI Press.
Bjorklund, D.F., Ornstein, P. A., & Haig, J.R. (1977).
Developmental differences in organization and recall:
Training in the use of organizational techniques.
Developmental Psychology , 13 , 175-183.
Bjorklund, D.F. , & Thompson, B.E. (1983). Category typicality
effects in children's memory performance: Qualitative
and quantitative differences in the processing of
category information. Journal of Experimental Child
Psychology , 35 , 329-344.
Bjorklund, D.F., & Zeman, B.R. (1982). Children's
organization and metamemory awareness in their recall
of familiar information. Child Development , 53 , 799-810.
Bjorklund, D.F. , & Zeman, B.R. (1983). The development of
organizational strategies in children's recall of
familiar information: Using social organization to
recall the names of classmates. International Journal
of Behavioral Development , 6_, 341-353.
Buchanan, J.J., & Bjorklund, D.F. (1988, March).
Developmental differences in the acquisition and
transfer of a memory strategy . Paper presented at the
Conference on Human Development, Charleston, SC.
Bull, J., & Farrand, J. (1977). The Audubon Society field
guide to North American Birds : Eastern region . New
York: Alfred A. Knopf.
Chi, M.T.H. (1978). Knowledge structures and memory
development. In R. Siegler (Ed.), Children's thinking :
What develops ? (pp. 73-96). Hillsdale, NJ: Erlbaum.
Chi, M.T.H. (1985). Interactive roles of knowledge and
strategies in development. In S. Chapman, J. Segal, & R.
Glaser (Eds.), Thinking and learning skills : Current
research and open questions (Vol. 2, pp. 457-483).
Hillsdale, NJ: Erlbaum.
Chi, M.T.H., & Ceci, S.J. (1987). Content knowledge: Its
role, representation, and restructuring in memory
development. In H.W Reese (Ed.), Advances in child
development and behavior (Vol. 20, pp. 91-142).
Orlando, FL: Academic Press.
Chi, M.T.H., & Koeske, R.D. (1983). Network representation of
a child's dinosaur knowledge. Developmental Psychology ,
Collins, A.M., & Quillian, M.R. (1969). Retrieval time from
semantic memory. Journal of Verbal Learning and Verbal
Behavior , 8, 240-247.
Corsale, K., & Ornstein, P. A. (1980). Developmental changes
in children's use of semantic information in recall.
Journal of Experimental Child Psychology , 3_0_' 231-245.
Daehler, M.W., & Greco, C. (1985). Memory in very young
children. In M. Pressley & C.J. Brainerd (Eds.),
Cognitive learning and memory in children : Progress in
cognitive development research (pp. 49-79). New York:
DeLoache, J.S., Cassidy, D.J., & Brown, A.L. (1985).
Precursors of mnemonic strategies in very young
children's memory. Child Development , 56 , 125-137.
DeMarie-Dreblow, D. (1988, April). Which type of knowledge
predicts recall : Category-generated listing, fact-
generation, or word-category matching ? Paper presented
at the annual meeting of the Midwestern Psychological
Association, Chicago, IL.
DeMarie-Dreblow, D., & Miller, P.H. (1988). The development
of children's strategies for selective attention:
Evidence for a transitional period. Child Development ,
59 , 1159-1187.
Dempster, F.N. (1981). Memory span: Source of individual and
developmental differences. Psychological Bulletin , 89 ,
Dempster, F.N. (1985). Short-term memory development in
childhood and adolescence. In C.J. Brainerd & M.
Pressley, Basic processes in memory development :
Progress in cognitive development research (pp. 209-
248). New York: Springer-Verlag .
Ebbinghaus, H. (1913). Memory . New York: Columbia University
Flavell, J.H. (1970). Developmental studies of mediated
memory. In H.W. Reese & L.P. Lipsitt (Eds.), Advances
in child development and behavior (Vol. 5, pp. 181-
211). New York: Academic Press.
Flavell, J.H., Beach, D.H., & Chinsky, J.M. (1966).
Spontaneous verbal rehearsal in a memory task as a
function of age. Child Development , 37 , 283-299.
Frankel, M.T., & Rollins, H.A. (1982). Age-related
differences in clustering: A new approach. Journal of
Experimental Child Psychology , 34 , 113-122.
Gobbo, C, & Chi, M. (1986). How knowledge is structured and
used by expert and novice children. Cognitive
Development , 1, 221-237.
Kee, D.W., & Bell, T.S. (1981). The development of
organizational strategies in the storage and retrieval
of categorical items in free-recall learning. Child
Development , 52 , 1163-1171.
Lange, G. (1973). The development of conceptual and rote
recall skills among school age children. Journal of
Experimental Child Psychology , 15 , 394-406.
Lange, G.W. ( 1978 ) .Organization-related processes in
children's recall. In P. A. Ornstein (Ed.), Memory
development in children (pp. 101-128). Hillsdale, NJ:
Lindberg, M.A. (1980). Is knowledge base development a
necessary and sufficient condition for memory
development? Journal of Experimental Child Psychology ,
Miller, P.H., 6. Harris, Y.R. (in press). Preschoolers'
strategies of attention on a same-different task.
Developmental Psychology .
Miller, P.H., Haynes, V.F., DeMarie-Dreblow, D., & Woody-
Ramsey, J. (1986). Children's strategies for gathering
information in three tasks. Child Development , 57 , 1429-
Moely, B.E., & Jeffrey, W.E. (1974). The effect of
organization training on children's free recall of
category items. Child Development , 45 , 135-143.
Moely, B.E., Olson, F.A., Halwes, T.G. , & Flavell, J.H.
(1969). Production deficiency in young children's
clustered recall. Developmental Psychology , 1, 26-34.
Moely, B.E., & Shapiro, S.I. (1971). Free recall and
clustering at four age levels: Effects of learning to
learn and presentation method. Developmental
Psychology , 4, 490.
Naus, M.J., Ornstein, P. A., & Aivano, S. (1977).
Developmental changes in memory: The effects of
processing time and rehearsal instructions. Journal of
Experimental Child Psychology , 23 , 237-251.
Olsen, G.M. (1983). Discussion: The past ten years. In M.T.H.
Chi ( Ed . ) , Trends in memory development research ( pp .
108-115). Basel: Karger.
Ornstein, P. A., Baker-Ward, L., & Naus, M.J. (in press). The
development of mnemonic skill. In F.E. Weinert & M.
Perlmutter (Eds.), Memory development: Universal changes
and individual differences . Hillsdale, NJ: Lawrence
Ornstein, P. A., & Naus, M.J. (1985). Effects of the knowledge
base on children's memory strategies. In H.W Reese
( Ed -), Advances in child development and behavior (Vol
19, pp. 113-148). New York: Academic Press.
Ornstein, P. A., Naus, M.J., & Liberty, C. (1975). Rehearsal
and organizational processes in children's memory.
Child Development , 46 , 818-830.
Ornstein, P. A., Naus, M.J., & Stone, B.P. (1977). Rehearsal
training and developmental differences in memory
Developmental Psychology , i_3, 15-24.
Pascual-Leone, J. (1970). A mathematical model for the
transition rule in Piaget's developmental stages. Acta
Psycholoqica . 32, 301-345.
Peterson, R.T. (1947). A field guide to the birds : Eastern
land and water birds . Boston: Houghton Mifflin~
Pressley, M. (1982). Elaboration and memory development.
Child Development , 53 , 296-309.
Pressley, M., Borkowski, J.G., & O'Sullivan, J. (1985).
Children's metamemory and the teaching of memory
strategies. In D.L Forrest-Pressley , G.E. MacKinnon, &
T.G. Waller (Eds.), Metacognition, cognition, and human
performance (pp. 111-153). Orlando, FL: Academic Press.
Pressley, M., & Ross, K.A. (1984). The role of strategy
utility knowledge in children's strategy decision
making. Journal of Experimenta l Child Psychology 38
491-504. z ax ' — '
Rabinowitz, M. (1984). The use of categorical organization:
Not an all-or-none situation. Journal of Exp erimental
Child Psychology , 38, 338-351.
Rao, N & Moely, B.E. (1987, April). Training mainten ance
and generali zation of an organizational stra tegy. Pan^r
presented at the biennial meeting of the Society for
Research in Child Development, Baltimore, MD.
Richman, C.L., Nida, S . , & Pittman, L. (1976). Effects of
meamngfulness on child free-recall learning
Developmental Psychology , 12, 460-465.
Robbins, C.S., Bruun, B., & zim, U.S. (1966). Birds o f North
America. New York: Golden Press. : ■ ±I± -
Roenker, D.L., Thompson, C.P., & Brown, S.C. (1971)
Comparison of measures for the estimation of clustering
in free recall. Psychological Bulletin , 76 , 45-48.
Saarnio, D.A. (1987). Knowing, remembering, and developing :
What are the relationships ? Unpublished manuscript
available from the author, Northern Illinois University,
Steward, L. (Ed.). (1977). The illustrated encyclopedia of
birds. Prague, Czechoslovakia: Artia.
Trabasso, T. (1983). Discussion: What is memory to be the
development of? In M.T.H. Chi (Ed.), Trends in memory
development research (pp. 116-122). Basel: Karger.
Wellman, H.M. (in press). The early development of memory
strategies. In F.E. Weinert & M. Perlmutter (Eds.),
Memory development: Universal changes and individual
differences . Hillsdale, NJ: Lawrence Erlbaum.
Zembar, M.J., & Naus, M.J. (1985a, April). An argument for a
more integrative approach to the study of memory
development . Paper presented at the meeting of the
Southwestern Psychological Association, Austin, TX.
Zembar, M.J., & Naus, M.J. (1985b, April). The combined
effects of knowledge base and mnemonic strategies on
children's memory . Paper presented at the biennial
meeting of the Society for Research in Child
Development, Toronto, Canada.
Zim, H.S., & Gabrielson, I.N. (1956). Birds : A guide to the
most familiar American birds. New York: Golden Press.
Darlene DeMarie Dreblow was born in New York, but she
never considered herself a "New Yorker." She earned a BA
from Marietta College in 1974, an MEd from Ohio University
in 1978, an MS from the University of Florida in 1985, and
will be receiving her PhD from the University of Florida in
1988, but she currently is not a student. She loved teaching
elementary school children (looking forward to every day when
she taught first grade, second grade, and learning
disabilities for seven years) but left to attend graduate
school anyway. After striving to achieve a "mellow"
existence for many years, Darlene broke the life-events
record during graduate school.
Darlene hates the cold but currently lives in
Cambridge, Ohio. She lives on a 98-acre farm, but she is not
a "farmer." She teaches at a small, liberal arts college
(Muskingum College in New Concord, Ohio), but she plans to
remain active in research. She has a family but rarely has
time to be a "mother" or a "wife."
These ironies in life have been inevitable. Hopefully,
they have strengthened Darlene for her pursuit for harmonious
relations among self, soul, and situation in the future.
o D inion C ?? tifY ^ that J have read this stud Y and that in my
opinion it conforms to acceptable standards of scholarlv
Patricia H. Miller, Chairperson
Professor of Psychology
-i^ 1 ce ftify that I have read this study and that in mv
SeSen^M cont ™* to acceptable standards of scholar^
Son" n^i-H 111 ^ adequate, in soope and quality, as
a Dissertation for the degree of Doctor of Philosophy.
Professor of Psychology
Professor of Psychology
Richard Griggs / /
Professor of Psychology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
ft. a ™~-4J± Qvp i a a.£L
J&Vnes Algina Y
P^Jafessor of Foundations of
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Assistant Professor of Psychology
This dissertation was submitted to the Graduate Faculty
of the Department of Psychology in the College of Liberal
Arts and Sciences and to the Graduate School and was accepted
as partial fulfillment of the requirements for the degree of
Doctor of Philosophy.
December, 1988 Dean, Graduate School